Oxidation of Disinfection Byproducts and
Algae-related Odorants by UV/H2O2
Chang Hyun Jo
Dissertation submitted to the faculty of the Virginia Polytechnic Institute and State University in partial fulfillment of the requirements for the degree of
Doctor of Philosophy In Civil and Environmental Engineering
Dr. Andrea M. Dietrich, committee chair Dr. John T. Novak, committee member Dr. John C. Little, committee member
Dr. Marc A. Edwards, committee member Dr. James M. Tanko, committee member Dr. Susan E. Duncan, committee member
August 26, 2008 Blacksburg, Virginia
UV/H2O2, Odorant, Disinfection byproduct, Advanced oxidation process
Copyright © 2008, Chang Hyun Jo
Oxidation of Disinfection Byproducts and Algae-related
Odorants by UV/H2O2
Chang Hyun Jo
ABSTRACT This research involved an investigation of the application and reaction mechanisms of
UV/H2O2 for the simultaneous removal of regulated halogenated disinfection byproducts
(DBPs) and odorous aldehydic algal byproducts in the presence of geosmin and 2-
methylisoborneol, which are earthy-musty odorants that commonly occur in drinking water.
UV/H2O2 is an expensive advanced oxidation process that is used to successfully control
geosmin and 2-methylisoborneol. The aqueous oxidation of odorous aldehydes and
halogenated DPBs were compared to that of the earthy-musty odorants and the changes to the
sensory properties of the drinking water were examined. Geosmin, 2-methylisoborneol,
heptadienal, decadienal, and nonadienal, hexanal, and the two most prevalent classes of DBPs,
trihalomethanes (THMs) and haloacetic acids (HAAs) were oxidized by UV photolysis alone
and the UV/H2O2 process with 6 mg/L H2O2 and realistic ng/l to µg/L concentrations of the
test compounds.
The di-, and tri-brominated THMs and HAAs were substantially (80-99%) removed by
direct UV photolysis mechanism at the same UV/H2O2 dose required for removing 95% of
geosmin and 65% of 2-methylisoborneol with faster reaction rates for the more bromine
substituted compounds. The C-Br bond cleavage is the first step of brominated HAAs
degradation by UV photolysis, and followed by either of two second steps: reaction with
oxygen producing peroxyl radical or interaction with water molecule causing O-H
insertion/H-Br elimination.
Trichloromethane and mono-, di-, and tri-chlorinated HAAs were not substantially removed
under the same conditions used for the brominated compounds. The principal removal
mechanism was by the reaction with hydroxyl radical for the UV/H2O2 process. The second
order reaction rate constants were on the order of 106 - 108 M-1 s-1 with faster reaction rates
iii
for the less chlorine substituted compounds. Based on the reaction rates, hydrogen and
halogen ion balance, and isotope effect, both hydrogen abstraction and electron transfer
reaction were involved in the first steps of the chlorinated HAA degradation.
Three odorous aldehydes - heptadienal, decadienal, and nonadienal - were removed faster
than geosmin or 2-methylisoborneol, and direct UV photolysis was the principal reaction
mechanism for the removal of these unsaturated aldehydes. Hexanal was poorly removed. In
sensory tests, new odors such as sweet or chalky odors were produced while the
concentration and initial odor intensity of these fishy/grassy-smelling aldehydes were reduced
with increasing exposure time to UV/H2O2. Carbonyl compounds were detected as products
of the UV photolysis of nonadienal. These carbonyls were not removed by further UV
irradiation, which was thought to be partially related with production of new odors.
The results indicate that the UV/H2O2 is effective to control both odorous compounds and
brominated DBPs. This process can be seasonally applied to control both contaminants
especially, in the warm summer when both odorants and DBPs have their higher
concentrations. Removal of brominated DBPs can be a significant addition to water utilities
that have difficulty in meeting regulatory levels for these highly toxic compounds. The
result on the removal of odorous aldehydes indicate that new types of odors were produced
from the oxidation of odorous aldehydes suggesting sensory test coupled with chemical
analysis should be considered in designing oxidation process to control recalcitrant odorants.
iv
Table of Contents
Abstract ...........................................................................................................ⅱ
Table of Contents ...........................................................................................ⅳ
List of Tables .................................................................................................ⅵ
List of Figures ................................................................................................ⅶ
Acknowledgement .........................................................................................ⅸ
Chapter 1. Introduction ....................................................................................1
Chapter 2. Review of Literature .......................................................................5
1. General concepts of UV application for drinking water.................................................. 5
2. Fundamentals of AOPs .................................................................................................. 7
3. Fundamentals of UV/H2O2 ............................................................................................ 9
4. DBP(FP) removal by AOPs ......................................................................................... 16
5. Taste/odor and AOP..................................................................................................... 18
6. Kinetics of geosmin/ 2-MIB and DBPs with hydroxyl radical...................................... 22
7. Reaction mechanism of DBPs and Geosmin/2-MIB in UV/H2O2 ................................. 23
Chapter 3. Simultaneous Removal of Odorants and Disinfection Byproducts by
UV/H2O2 Advanced Oxidation Process .......................................................... 30
Introduction ................................................................................................................... 31
Materials and Methods ................................................................................................... 35
Results ........................................................................................................................... 37
Discusstion .................................................................................................................... 45
Conclusion ..................................................................................................................... 45
v
Chapter 4. Reaction Mechanism of Haloacetic acid in UV/H2O2 Advanced
Oxidation Process .......................................................................................... 47
Introduction ................................................................................................................... 48
Materials and Methods ................................................................................................... 50
Results ........................................................................................................................... 52
Discussion ..................................................................................................................... 67
Conclusion ..................................................................................................................... 69
Chapter 5. Removal of Odorous Aldehydes by UV/H2O2 ............................... 70
Introduction ................................................................................................................... 70
Materials and Methods ................................................................................................... 71
Results ........................................................................................................................... 73
Discusstion .................................................................................................................... 78
Conclusion ..................................................................................................................... 79
References ...................................................................................................................... 80
vi
List of Tables
Table 1-1. Odorants and DBPs selected for this research ....................................................... 4
Table 2-1. Characteristics of AOPs ...................................................................................... 7
Table 2-2. Reduction potential of oxidants ......................................................................... 10
Table 2-3. Second order rate constants of DBPs and odorants with hydroxyl radical............ 22
Table 3-1. Typical concentrations of compounds in the research.......................................... 36
Table 3-2. Comparison of % removal in de-ionized water and reference water .................... 44
Table 4-1. concentrations of HAA compounds examined..................................................... 51
Table 4-2. Apparent pseudo-first order reaction rate constants for UV photolysis of three
brominated HAAs .............................................................................................................. 54
Table 4-3. Comparison of measured and calculated ∆[H+]/∆HAA based on percent
mineralization ..................................................................................................................... 59
Table 4-4. Second order reaction rate constants of chlorinated HAAs ................................ 61
Table 4-5. Comparison of measured and expected parameters of chlorinated HAAs ............ 67
Table 5-1. Odorants selected for this research...................................................................... 72
vii
List of Figures
Figure 1-1. Paradigm shift in drinking water quality.............................................................. 2
Figure 2-1. The drinking water taste and odor wheel .......................................................... 18
Figure 3-1. UV irradiation system and quartz reactor........................................................... 35
Figure 3-2. Molar absorption coefficients measured at 254 nm in this research.................... 38
Figure 3-3. Comparison of removal rate between UV photolysis and UV/H2O2 for geosmin
and 2-MIB .......................................................................................................................... 39
Figure 3-4. Removal of geosmin/2-MIB and THMs with UV/H2O2..................................... 40
Figure 3-5. Comparison of removal rates between UV photolysis and UV/H2O2 for
brominated THMs............................................................................................................. 41
Figure 3-6. Removal rates of halogenated methanes measured for individual compounds .. 42
Figure 3-7. Removal of Geosmin/2-MIB and HAAs with UV/H2O2 .................................... 43
Figure 3-8. Comparison of removal rates between UV photolysis and UV/H2O2 for
brominated HAAs ............................................................................................................. 44
Figure 4-1. Removal rate of brominated HAAs and tribromomethane by UV photolysis ..... 53
Figure 4-2. Molar absorption coefficients measured at 254 nm in this research.................... 53
Figure 4-3. Molar increase of [H+] and [Br-] with molar decrease of three brominated HAAs
exposed to UV photolysis at 253.7 nm wavelength.............................................................. 55
Figure 4-4. Molar decrease of TOC with molar decrease of HAA concentration for three
brominated HAAs .............................................................................................................. 56
Figure 4-5. Removal rates of three chlorinated HAAs compared to trichloromethane .......... 60
Figure 4-6. Comparison of reaction rates with UV/H2O2 between deuterated MCAA and
viii
MCAA................................................................................................................................ 61
Figure 4-7. Transition state for hydrogen abstraction of DCAA; both chlorine atoms withdraw
electron density from the carbon atom................................................................................. 62
Figure 4-8. Transition state for hydrogen abstraction of MCAA .......................................... 62
Figure 4-9. Partial positive charge on the chlorinated carbon atom of acetate ion ............... 63
Figure 4-10. Molar increase of [H+] and [Cl-] compared to molar decrease of corresponding
chlorinated HAA................................................................................................................. 64
Figure 4-11. Molar decrease of TOC with molar decrease of three chlorinated HAAs ........ 64
Figure 5-1. Molar extinction coefficient measured in this research (M-1cm-1)....................... 73
Figure 5-2. Log removal of odorants with UV dose (6 mg/L H2O2 ) .................................... 74
Figure 5-3. Nonadienal concentration and odors as a function of UV dose (6 mg/L H2O2)... 75
Figure 5-4. Decadienal concentration and odors as a function of UV dose (6 mg/L H2O2) ... 75
Figure 5-5. Heptadienal concentration and odors as a function of UV dose (6 mg/L H2O2) .. 75
Figure 5-6. Hexanal concentration and odors as a function of UV dose (6 mg/L H2O2)........ 76
Figure 5-7. Comparison of PFBHA derivatized chromatograms for UV photolysis and
UV/H2O2 treatment of nonadienal ....................................................................................... 77
Figure 5-8. GC/MS chromatograms of PFBHA derivatized nonadienal ............................. 78
ix
Acknowledgement There was a turtle in America. What the turtle did was to keep going slowly without a long break because he couldn�t run
or fly. Many good people helped and supported the turtle. With their help, the turtle is about to finish his race. My wife, Sun Young and my son, Hyun Jae gave me a good reason I had to keep going. I
also want to express a deep appreciation to my parents. While I was in Blacksburg, I came to better realize how much family means to me. First of all, I�d like to greatly thank my advisor, Dr. Dietrich and her family. She was like
my aunt in America. She always shows me the way when I am lost, and supports me. I also thank my committee members (Dr. Novak, Dr. Little, Dr. Edwards, Dr. Duncan, Dr.
Tanko), department head, Dr. Knocke, and other professors for their good guidance. It was lucky for me to learn from them. Especially, I thank Dr. Tanko for helping me enter the radical chemistry world. My friends in our research group (Pinar, Andy, Jose, Ryan, Dave, Heather, Tim, and
Monique) supported and encouraged me a lot. I was happy to be with them. I also thank Betty, Beth, Jody, Julie, and other friends in my department for their kind help. My Korean friends, if there were not their help, I would have had much more difficult times
adapting myself to the life in America. My friend, Bruce, Nicki, and Angelo, I thank you guys for your friendship. That means a lot
to me. Last, I specially thank my company, Kwater for supporting me. I always thought how I was lucky to have a chance to study abroad even if it was a big challenge to me. The more I study science, the more I realize that humans just mimic what the mother nature
does. While I am studying, I also realized happiness lies rather in how we are related with each other than what we have or what we accomplished. I wish I had spent more time getting closer to my good friends in Blacksburg. I know this dissertation is just a minimum requirement for Ph.D and a first step to the
expertness. I might walk faster in my country than in America. However, I will remember that I was a turtle in a foreign country.
1
Chapter 1. Introduction
Drinking water treatment has evolved to fulfill demands for safe and clean water. At the
early 20th century, sanitary water treatment systems were required to inactivate pathogens and
supply a sufficient amount of water. Since then, many treatment techniques have been
introduced to the water industry in order to supply safe drinking water that is free of chemical
contaminants as well as biological contaminants, many of which were released into source
water as a result of civilization. However, there still have been concerns about the quality of
drinking water.
Most of the concerns about drinking water result from health issues. Disinfection
byproducts (DBPs) are one of the major health issues in the drinking water industry due to
their carcinogenicity and genotoxicity (Richardson, Plewa et al. 2007). Many utilities are
suffering from the disinfection byproducts problem, which is also frequently in conflict with
obtaining disinfection credit required to inactivate pathogens such as Giardia and
Cryptosporidium.
Currently, consumers require more than safe water, and more interest is being shown to
aesthetic issues such as taste and odor (Khiari 2004; Liang, Wang et al. 2007; Peter and Von
Gunten 2007). This trend indicates that consumers demand �more pleasant� or �more tasty�
water. Geosmin (trans-1,10-dimethyl-trans-9-decalol) and 2-MIB (2-methylisoborneol) are
typical earthy-musty smelling odorants found in surface water and subsequently, drinking
water. These compounds cause seasonal odor episodes, and are difficult to remove by
conventional water treatment processes, and easy for consumers to detect even at low
concentrations due to their low odor threshold levels (4-10 ng/L). Another widespread algae-
related odor problem is the fishy/grassy odor that is frequently produced from aldehyde
compounds. Aesthetic issues also frequently involve concerns about health issues, causing
consumer complaints because consumers tend to relate aesthetic issues to health risks.
Consequently, meeting the demands for taste and safety is the current agenda of the water
industry in 21st century (Figure 1-1). This research is a study on a treatment method,
UV/H2O2 advanced oxidation, which is being evaluated for removing odorous compounds
and disinfection byproducts and is known to be effective for disinfection.
2
Figure 1-1. Paradigm shift in drinking water quality
A variety of treatment processes have been developed and used to control taste and odor
compounds and DBPs, including activated carbon, ozonation, and advanced oxidation
process (AOP). AOP oxidizes contaminants with hydroxyl radical (·OH). AOP, like other
technologies developed by humans, basically mimics natural phenomena such as the
oxidation in the surface water or atmosphere by sunlight (Oppenlander 2003). AOP has an
advantage that it efficiently removes organic contaminants without production of residual
solids. Additionally, AOP, when it is combined with UV, is an alternative disinfection method
for pathogen inactivation (EPA 1999). UV/H2O2 is an AOP that has been applied to drinking water since the 1990s. In this process,
hydroxyl radicals are generated by the direct photolysis of H2O2 under UV irradiation (Liao
and Gurol 1995; Stefan, Hoy et al. 1996; Stefan and Bolton 1998; Stefan, Mack et al. 2000;
Rosenfeldt, Melcher et al. 2005; Rudra, Thacker et al. 2005; Xu, Gao et al. 2007). This
process has been known to efficiently remove organic contaminants, including recalcitrant
odorous compounds such as geosmin and 2-MIB, mainly by the hydroxyl radical reaction and
partially by direct UV photolysis (Beltran, Ovejero et al. 1993; Stefan, Hoy et al. 1996;
Stefan and Bolton 1998; Cater, Stefan et al. 2000; Stefan, Mack et al. 2000; Rosenfeldt,
Melcher et al. 2005; Rudra, Thacker et al. 2005; Paradis and Hoffman 2006; Rosenfeldt and
Linden 2007). AOPs are also thought to effectively remove other algae-related odorants such
as odorous aldehydes based on the measured second order reaction rate constant with
3
hydroxyl radical (Peter and Von Gunten 2007). However, less AOP research was performed
on the removal of other algae-related odorants than geosmin and 2-MIB. Furthermore, it was
reported that some algal metabolites were transformed into new types of odor by oxidation
(Dietrich, Hoen et al. 1995), and the fruity smelling aldehydes were produced from the
ozonation of drinking water (Anselme, Suffet et al. 1988; AWWARF 1995; Bruchet and
Duguet 2004). Therefore, further research is required to investigate how effectively algae-
related odorants can be removed, how odor descriptors change, and what types of new odors
are produced.
Recently, UV/H2O2 was applied to full scale water treatment plants (WTPs) to control
earthy-musty odors (geosmin and 2-MIB), N-nitrosodimethylamine (NDMA), and 1,4-
dioxane (Cotton and Collins 2006). Full scale UV/H2O2 systems utilize low intensity UV for
disinfection and high intensity UV for both disinfection and advanced oxidation (Cotton and
Collins 2006). The UV/H2O2 process is known to have several advantages compared to other
AOPs; simple operation procedure, small foot print, no regulated DBPs formation, and dual
mode (low intensity UV for disinfection, high intensity UV and H2O2 for advance oxidation)
(Legrini, Oliveros et al. 1993; Cotton and Collins 2006).
However, UV/H2O2, like other AOPs, typically cost much more than conventional
treatment. Total cost for applying UV/H2O2 to an existing 40 MGD utility with typical water
quality and taste/odor episode was estimated as $0.05-0.07/kgal in a field study (Royce and
Stefan 2005)[AMD1]. Due to the economical and practical aspects, AOP could be best applied
to address a seriously concerning contaminant or multiple contaminants. This research will
investigate DBP removal and its mechanism when UV/H2O2 is applied to control earthy-
musty odorous compounds. Additionally, the removal of algae-related odorous aldehydes by
UV/H2O2 and its effect on the sensory was studied. Geosmin and 2-MIB, and four types of
odorous aldehydes were used in this research as well as two most prevalent DBPs,
trihalomethanes (THMs) and haloacetic acids (HAAs) (Krasner, Weinberg et al. 2006) as
shown in Table 1-1. The objectives of this research were to investigate: 1) types of DBPs that
can be removed by UV/H2O2 dose for recalcitrant earthy-musty odor control, 2) mechanisms
involved in this DBP removal, 3) how effectively fishy/grassy smelling aldehydes are
removed, and 4) how odorous aldehydes are transformed after the advanced oxidation. This
research could be an addition to the AOP design that controls both taste/odor and DBP
problem.
4
Table 1-1. Odorants and DBPs selected for this research
Compounds Structure Guideline in
drinking water
Effect in drinking
water
trans-2,cis-6-
nonadienal
O
-
Cucumber/Fishy
Odor
trans-2,trans-4-
decadienal
O
-
Fishy/Oily/Cucumber
Odor
trans-2,trans-4-
heptadienal
O
-
Grassy/Oily/Fishy
Odor
Hexanal O
-
Grassy/Sweet
Odor
Geosmin
10 ng/L a Earthy odor
Odorants
2-MIB
10 ng/L a Musty odor
Trihalomethanes
(THMs)
C
H
X
Y
Z
X, Y, Z= Cl, Br, I
80 µg/L b Carcinogenic c
Disinfection
Byproducts Haloacetic acids
(HAAs)
C C
X O
O HY
Z
X, Y, Z= H,Cl, Br, I
60 µg/L b Genotoxic and
carcinogenic c
a Guideline in Korea and secondary standard in Japan b Maximum contaminant level in U.S c (Richardson, Plewa et al. 2007)
5
Chapter 2. Review of Literature
1. General concepts of UV application for drinking water
UV Irradiation
The UV spectrum can be classified as Vacuum UV (VUV, 100-200 nm), UV-C (200-
280nm), UV-B (280-315 nm), and UV-A (315-400 nm) based on wavelength. It is well
established that UV inactivates microorganisms by transforming DNA. In terms of germicidal
effects, the optimum UV range is between 245 and 285 nm because DNA does not absorb UV
above the wavelength of 300 nm (AWWA 1999; EPA 1999; Crittenden, Trussell et al. 2005).
UV is transmitted through water to be absorbed into or reflected off of the materials. No
residual is produced from the UV radiation, which is an advantage in terms of DBP formation.
However, a secondary chemical disinfectant is required to maintain a residual in the
distribution system (AWWA 1999; EPA 1999). UV demand of water, the absorption of energy
per unit depth or absorbance, can be measured by a spectrophotometer set at a wavelength of
254 nm. UV dose (fluence) can be represented as follows (EPA 1999):
D = I·t
D = UV dose (mJ/cm2 or mW·s/cm2)
I = Intensity (mW/cm2)
t = Exposure time (s)
Measurement of UV dose (fluence)
UV dose can be determined with the iodide/iodate actinometer by measuring triiodide ion
(I3-) produced from the UV photolysis of iodide ion (I-) at the wavelength of 352 nm. Iodate
ion (IO3-) plays a role of electron scavenger by inhibiting the reverse reaction of UV
photolysis (I· + e- → I- ). Reactions in this actinometry are as follows (Rahn 2004; Rahn,
Bolton et al. 2006):
I- + hν → I· + e-
2I· +2 I- → 2I2·-
2I2·- → I- + I3-
6
IO3- + e- + 2H2O → IO- + H2O2 +·OH +OH- (electron scavenging)
·OH + I- → I· + OH-
UV irradiance or incident intensity at a surface is typically measured by �collimated beam�
system that measures the intensity of collimated UV at the surface of the sample water
(Rosenfeldt, Melcher et al. 2005).
UV lamp
Three types of UV lamps are used in the water industry: (1) low-pressure, low-intensity
lamp, (2) low-pressure, high-intensity (high output) lamp, (3) medium-pressure, high
intensity lamp (Crittenden, Trussell et al. 2005). Both low-pressure and medium-pressure
lamps can be used for disinfection application. Low-pressure lamps have their maximum
energy output at a wavelength of 253.7nm, while the spectrum of medium pressure lamps
have energy output at wavelengths ranging from 180 to 1370 nm (EPA 1999; Crittenden,
Trussell et al. 2005). Fewer medium pressure lamps are required for an equivalent dosage
than low pressure lamps due to higher intensity. Several low-pressure lamps are
recommended compared to one medium pressure lamp for small systems because of
reliability of multiple lamps and cleaning cycle (EPA 1999).
UV dose required for pathogen inactivation
The UV dose required for effective inactivation is determined by site-specific data related
to the water quality and log removal requirements (EPA 1999). A UV dose of 36 mJ/cm2 was
required for 3-log inactivation of viruses (AWWA 1991; EPA 1999). Much higher dosages are
required for larger protozoa such as Cryptosporidium and Giardia inactivation (White 1992;
EPA 1999). To achieve 2-log inactivation of Giardia lamblia cysts, at least 121 mJ/cm2 was
required (Carlson 1982; EPA 1999). Since AOPs have been proven to be equal or more
effective than ozone for pathogen inactivation, UV used with ozone and H2O2 enhances the
disinfection effectiveness (EPA 1999).
7
2. Fundamentals of AOPs
Comparison of typical AOPs
Muller and Jekel compared three AOPs (UV/H2O2, O3/H2O2, O3/UV) in the pilot and full
scale study on the atrazine removal. Based on the comparison of electrical energy per order
(EEO), O3/H2O2 was reported to be the most economical process (Muller, Gottschalk et al.
2001; Muller and Jekel 2001). Characteristics of AOPs were compared in Table 2-1.
Table 2-1. Characteristics of AOPs (National Water Research Institute 2000)
AOP Major reaction Advantages Disadvantages
UV/H2O2 H2O2 + hν → 2·OH
- No bromate formation
- Can serve a disinfectant
- Full scale drinking water
treatments exist
- No off-gas treatment required
- No mass transfer between
Liquid and gas phase
- Interference of turbidity
- Interference of UV
absorbing compounds
O3/H2O2 H2O2 + H2O → HO2
- + H3O+
O3 + HO2- → ·OH + O2
- + O2
- The most economical process
based on EEO
- Efficient in MTBE treatment
- Established technology for
remediation
- Potential for bromate
formation
(controllable by O3/H2O2
ratio and pH)
- May require excessive H2O2
treatment
- May require off-gas
treatment
O3/UV O3 + H2O + hν → O2 + H2O2
H2O2 + hν → 2·OH
- More efficient at generating
·OH
- Energy and cost intensive
process
- May require off-gas
treatment
Ozonation
Basic reactions in ozonation
Ozone itself is a selective oxidant but it oxidizes organic compounds through hydroxyl
8
radical produced from the reaction of ozone and natural organic matter (NOM) or auto-
decomposition as below (AWWA 1999; Ho, Newcombe et al. 2002; Ho, Croue et al. 2004;
Westerhoff, Nalinakumari et al. 2006). Higher pH is recommended for ozonation to produce
more hydroxyl radical because ozone is dominant at lower pHs.
O3+ NOM → ·OH + other products
O3+ OH- → ·HO2 + ·O2-
·HO2 ↔ H+ + ·O2- (pKa = 4.8)
O3+·O2- + H2O → ·OH + OH- + 2O2
O3+·OH → ·HO2 + O2
Rct in ozonation process
In the ozonation process, compounds are oxidized by hydroxyl radical rather than ozone
itself because hydroxyl radical is highly reactive and nonspecific while ozone itself is a
selective oxidant to many organic compounds (Michael and Von Gunten 1999). Rct is the
parameter for determining hydroxyl radical concentration based on ozone concentration. In
this concept, a change in concentration of a ·OH-probe compound, para-chlorobenzoic acid
(pCBA) is measured and equated to a hydroxyl radical concentration that is difficult to
directly measure. Rct is specific to given water quality, and can be calculated from the pCBA
removal and dissolved ozone concentrations. Dissolved ozone concentration can be measured
by Indigo method where decreased indigo trisulfonic acid concentration by ozone is
measured by the decreased light absorption at 600 nm (Bader and Hoign 1981). Rct is useful
for determining the steady state hydroxyl radical concentration by measured ozone
concentration.
Because Rct is constant for given water quality and independent of the reaction time,
hydroxyl radical concentration can be represented as follows:
9
3. Fundamentals of UV/H2O2
UV/H2O2 process
The UV/H2O2 process is a homogeneous AOP in which hydroxyl radicals are generated by
the direct photolysis of H2O2 under UV irradiation and radical chain reactions (Liao and
Gurol 1995; Stefan, Hoy et al. 1996; Stefan and Bolton 1998; Stefan, Mack et al. 2000;
Rosenfeldt, Melcher et al. 2005; Rudra, Thacker et al. 2005; Xu, Gao et al. 2007).
H2O2 + hν → 2OH· Light absorption/initiation ·OH + H2O2 → HO2· + H2O Propagation HO2· + H2O2 → ·OH + H2O + O2 HO2· + HO2· → H2O2 + O2 Termination
The quantum yield for this reaction, which is the number of moles of H2O2 decreased per
mole of photon absorbed, has been reported as 1.0 for the overall quantum yield (ФT), and as
0.5 for the primary quantum yield (ФP). (Liao and Gurol 1995; Oppenlander 2003). In the
case of the hydroxyl radical reaction, a steady state radical concentration is assumed due to
relatively higher H2O2 concentration (mg/L level) than contaminants concentration
(ng/L~µg/L level) (Sharpless and Linden 2003; Rosenfeldt, Melcher et al. 2005; Pereira,
Weinberg et al. 2007; Xu, Gao et al. 2007).
Advantages and limits of UV/H2O2 process
The UV/H2O2 process has a number of advantages compared to other AOPs: commercial
availability of the oxidant, thermal stability, on-site storage, infinite solubility in water, no
mass transfer problems between two phases, minimal capital investment, simple operation
procedure, small foot print, no regulated DBPs formation, dual mode (low intensity UV for
disinfection, high intensity UV and H2O2 for advance oxidation) (Legrini, Oliveros et al.
1993; Cotton and Collins 2006).
However, H2O2 has a relatively small absorption cross section, the ability to absorb a
photon of a particular wavelength; this limits the rate of hydroxyl radical formation.
Therefore, in order to obtain higher rate of hydroxyl radical formation, Xe-doped Hg arc
lamp that has a strong emission at 210-240 nm wavelength is used. At this range of
wavelength, H2O2 has a higher molecular extinction coefficient. As in all AOPs, hydroxyl
radical is trapped by scavengers, such as bicarbonate and carbonate in water, which is the
10
main disadvantage of UV/H2O2 process (Legrini, Oliveros et al. 1993). Oxidation by hydroxyl radical
Hydroxyl radical is a very strong oxidants based on reduction potential shown in Table 2-2.
Table 2-2. Reduction potential of oxidants (AWWA 1999)
Species Reduction potential, E0red (V)
Hydroxyl radical
Atomic oxygen
Ozone
Hydrogen peroxide
Permanganate
Chlorine dioxide
Chlorine
Molecular Oxygen
2.80
2.42
2.07
1.78
1.68
1.57
1.36
1.23
Hydroxyl radical oxidizes organic compounds generally by hydrogen abstraction generating
organic radicals. The organic radical yields peroxyl radical by the reaction with oxygen.
These organic and peroxyl radicals initiate oxidative chain reactions leading to mineralization
where the final products are carbon dioxide, water, and inorganic salts. Another mechanism is
electron transfer to hydroxyl radicals leading to hydroxyl ion (Legrini, Oliveros et al. 1993).
HO· + RH → R· + H2O hydrogen abstraction
R· + O2 → RO2· peroxyl radical production
HO· + RX → RX·+ + HO- electron transfer to hydroxyl radical
The rate and efficiency of the oxidation process performed by hydroxyl radical depends on
the energy required to homolyze a given chemical bond, and the concentration of dissolved
oxygen (Legrini, Oliveros et al. 1993). Based on the characteristics of water, other types of
radicals can react with organic pollutants. Superoxide radical (HO2·), carbonate radical(CO3·-
/HCO3·), or phosphate radicals (HPO4·-) can oxidize organic contaminants (Crittenden, Hu et
al. 1999).
ROH,UV parameter in UV/H2O2 process
Recently, the ROH,UV concept, which is similar to Rct concept in ozonation, was indroduced
11
to characterize the water-specific effectiveness of the UV/H2O2. ROH,UV is defined as the
hydroxyl radical exposure per UV fluence, and affected by scavenging and UV absorbance of
water matrix (Rosenfeldt and Linden 2007). From the ROH,UV and UV fluence, hydroxyl
radical concentration produced from the UV/H2O2 reaction can be determined.
Both sides are divided by average UV fluence rate E0 (mW/cm2) to convert a time-based rate
constant into a fluence-based constant yielding following equation.
Direct UV photolysis
Photoxidation generally takes place in two ways. First is by the excitation of an organic
substrate followed by the electron transfer from the excited state to the ground state
molecular oxygen (eq. 1 and 2). Second is by the homolysis where radicals are formed
followed by the subsequent reaction with oxygen (eq. 3 and 4) (Legrini, Oliveros et al. 1993).
In order to absorb UV, a compound has to possess a UV absorbing chromophore at 253.7nm
for low pressure UV lamp or at wider range of wavelength for medium pressure UV lamp.
hν
C → C* (1)
C* + O2 → C·+ + O2·- (2)
hν
12
R-X → R· + X· (3)
R· + O2 → RO2· (4)
In many UV/H2O2 studies, direct UV photolysis, without H2O2, also has been shown to
contribute to removal of organic compounds. However, removal rates depended on the type
of compounds. Geosmin and 2-MIB were removed 40% and 20% respectively at the UV
irradiance of 1,700 mJ/cm2 (Rosenfeldt, Melcher et al. 2005). Diazinon decreased 20% at the
UV irradiance of 600 mJ/cm2 (Shemer and Linden 2006). Microcystin decreased 50% at the
UV irradiance of about 3,000mJ/cm2 (Qiao, Li et al. 2005). In regard to UV photolysis of
DBPs, it was reported that brominated THMs were photolysed and the quantum yield of the
photolysis was 0.43 (Nicole, De Laat et al. 1991). In the same research, more bromine
substituted THMs were shown to be photolysed faster. In a study of chlorinated swimming
pool water, tribromomethane and chlorodibromomethane levels were reported to decrease
significantly with UV irradiation of 145mJ/cm2 (Cassan, Mercier et al. 2006). For the direct
UV photolysis and hydroxyl radical reactions of organic compounds such as geosmin/2-MIB,
diethyl phthalate, and pharmaceutical compounds in UV/H2O2, pseudo-first order reaction
rate models at a wavelength (λ) were proposed as follows (Sharpless and Linden 2003;
Rosenfeldt, Melcher et al. 2005; Pereira, Weinberg et al. 2007; Xu, Gao et al. 2007):
[ ] ' [ ]dd C k C
dt− =
' dk = the measured pseudo-first order rate constant of direct photolysis (s-1)
= , ( )S Ck λ ( )c λΦ , ( )S Ck λ = specific rate of UV absorption by the compound (Es·mol-1s-1)
= 0 ( )( ) ( )[1 10 ]
( )
a zpE
a z
λλ ε λλ
−−
( )C λΦ = quantum yield of compound (mol Es-1)
0 ( )pE λ = incident photon irradiance (mEscm-2s-1)
( )ε λ = molar extinction coefficient of compound at a specific wavelength (M-1cm-1)
( )a λ = solution absorbance at a specific wavelength (cm-1)
z = solution depth (cm)
Kinetics in UV/H2O2 process
For the direct UV photolysis and hydroxyl radical reactions of organic compounds such as
geosmin/2-MIB, diethyl phthalate, and pharmaceutical compounds in UV/H2O2, pseudo-first
13
order reaction models at a wavelength (λ) were proposed as follows (Sharpless and Linden
2003; Rosenfeldt, Melcher et al. 2005; Pereira, Weinberg et al. 2007; Xu, Gao et al. 2007):
[ ] '[ ]d C k Cdt
− =
Where, ' ' 'd ik k k= +
'k = the observed pseudo-first order rate constant (s-1)
' dk = the measured pseudo-first order rate constant of direct photolysis (s-1)
= ( )Sk λ ( )c λΦ , ( )S Ck λ = specific rate of UV absorption by the compound (Es·mol-1s-1)
= 0 ( )( ) ( )[1 10 ]
( )
a zpE
a z
λλ ε λλ
−−
( )C λΦ = quantum yield of compound (mol· Es-1)
0 ( )pE λ = incident photon irradiance (mEs·cm-2s-1)
( )ε λ = molar extinction coefficient of compound at a specific wavelength (M-1cm-1)
( )a λ = solution absorbance at a specific wavelength (cm-1)
z = solution depth (cm)
' ik = the measured pseudo-first order rate constant of the reaction with ·OH
/' [ ]i C OH ssk k OH=
/C OHk = Second order reaction rate constant of compound and ·OH
[ ]ssOH = steady state ·OH concentration
In case of the hydroxyl radical reaction, a steady state hydroxyl radical concentration is
assumed due to relatively higher H2O2 concentration (2-30 mg/L) (Sharpless and Linden
2003; Rosenfeldt, Melcher et al. 2005; Cotton and Collins 2006; Pereira, Weinberg et al.
2007; Xu, Gao et al. 2007).
2 2, 2 2
,
( ) ( )[ ][ ]
[ ]S H O OH
ssS OH i
i
k H OOH
k Sλ λΦ
=∑∑
2 2, ( )S H Ok λ = specific rate of UV absorption by H2O2 (Es mol -1s-1)
( )OH λΦ = quantum yield for ·OH formation (≈ 1mol Es-1)
,S OHk
= second order reaction rate constant of scavenging species and ·OH (M-1s-1)
[ ]iS = concentration of scavenging species (M)
14
Factors affecting photochemical AOP performance
UV transmittance (UVT) affects dose, which is related to hydroxyl radical formation.
Carbonate and bicarbonate are the most common inorganic hydroxyl radical scavenger in
natural water (Crittenden, Hu et al. 1999; Cotton and Collins 2006). Although these two
scavengers produce carbonate radicals (·CO3-/·HCO3) as shown below, which can react with
organic contaminants, these reactions are not significant (Crittenden, Hu et al. 1999).
CO32- + ·OH → ·CO3
- + OH-
HCO3- + ·OH → ·HCO3 + OH-
Bromide and chloride ions are also known to scavenge hydroxyl radical with the reaction
rate constants of 1010 M-1s-1 and 2 x 107 M-1s-1, respectively as shown below (von Gunten and
Hoigne 1994; Nakatani, Hashimoto et al. 2007). Chloride (Cl-) was shown to substantially
contribute to scavenging of hydroxyl radical in typical drinking water based on the rate
constant for the reaction with hydroxyl radical (2 x 107 M-1s-1) and its typical concentration
(Nakatani, Hashimoto et al. 2007).
Br- + ·OH → BrOH-
Cl- + ·OH → ClOH-
Natural organic matters (NOM), such as humic or fulvic substance, lower the efficiency of
AOPs by absorbing UV and scavenging hydroxyl radicals (Crittenden, Hu et al. 1999). NOM
with higher UV absorbing properties consumed ozone and produced hydroxyl radical at a
higher rate in the ozonation process (Ho, Croue et al. 2004). In general, maximum UV/H2O2
performance can be obtained in slightly acidic condition. UV/H2O2 performance decreases
with suspended solids, nitrate and iron concentrations (TrojanUV 2003). The optimum H2O2
dose in UV/H2O2 process is required to be determined because an excessive dose can reduce
the oxidation rate (Wang, Hsieh et al. 2000) by scavenging hydroxyl radical and producing
less reactive hydroperoxyl radical as below (Legrini, Oliveros et al. 1993).
HO· + H2O2 → H2O + HO2·
Quenching of the peroxide residual is required after it passes through the UV reactor in full
scale process. In pilot or full scale plant, chlorine is used to quench H2O2 (Royce and Stefan
2005). Since oxidation potential of H2O2 (1.77V) is greater than Cl2 (1.36V) or OCl- (0.89V),
H2O2 can be quenched by free chlorine as follows (Batterman, Zhang et al. 2000; Liu,
Andrews et al. 2003):
15
Cl2 + H2O2 → 2Cl- + 2H+ + O2
EEO (Electrical Energy per Order) (Bolton and Stefan 2002; TrojanUV 2003; Rosenfeldt,
Melcher et al. 2005)
EEO (Electrical Energy per Order) is the metric for measuring efficiency of the UV
oxidation process, and has been used in industrial applications. EEO is defined as the electrical
energy required for reducing the contaminant concentration by one order of magnitude (1-log
or 90%) per cubic meter or 1000 gallons of water as follows:
0.06 = conversion factor of time and volume
C0= intial (influent) concentration
C = final (effluent) concentration
EEO is specific for the reactor type, contaminants, and water quality. The less EEO means that
the lower power is required by the system. Parameters affecting EEO are 1) reactor design , 2)
lamp type, 3) water quality such as UV transmittance (UVT) and scavengers concentration, 4)
lamp age, 5) flow rate, 6) hydrogen peroxide concentration, and 7) contaminant
characteristics such as quantum yield, molar extinction coefficient, and hydroxyl radical
reaction rate. Industries are using this EEO as a comprehensive parameter of UV oxidation
performance because UV dose should be normalized by other chemical dose such as
hydrogen peroxide dose, and there is a nonlinear relationship between power draw and UV
dose. For the removal of geosmin and 2-MIB by medium pressure (MP) UV and hydrogen
peroxide, EEOs of 0.5 to 1.2 was reported by researchers (Cotton and Collins 2006).
Application to full scale WTPs
Recently, UV/H2O2 has been applied in several full scale water treatment plants (WTPs) to
mainly control earthy-musty odor (geosmin and 2-MIB), N-nitrosodimethylamine (NDMA),
and 1,4-dioxane (Cotton and Collins 2006). In 2004, PWN water supply company in Holland
applied UV/H2O2 for disinfection and reducing organic pollutants with removing breakpoint
chlorination. O3/H2O2 was also proved to be a good process for this case but was not selected
due to high bromated levels formed by the process (Martijn, Kruithof et al. 2006). The
optimum dose range of hydrogen peroxide for removing geosmin and 2-MIB was reported to
16
be 6-10 mg/L in pilot and bench scale research (Cotton and Collins 2006; Paradis and
Hoffman 2006).
The cost of UV/H2O2
Based on EEO and electricity cost ($0.075 per kWh), a UV/H2O2 treatment cost of $0.35 per
1,000 gallons to remove the geosmin and 2-MIB by one order of magnitude was reported
(Rosenfeldt, Melcher et al. 2005). Conceptual level capital, operation and management cost
were estimated in a research. When design flow rate is 50 MGD (≈190,000 m3/d), taste and
odor removal is 90%, hydrogen peroxide dose is 10 mg/L, and UVT is 90%, total capital cost
and annual operation/management cost were calculated to be about $17,000,000 and
$1,000,000, respectively (Cotton and Collins 2006). Total cost including installment capital
and operation and management cost for a 40 MGD utility with typical water quality and
taste/odor episode was estimated as $0.05-0.07/kgal in a field study while ozonation cost was
estimated as $0.06-0.09/kgal (Royce and Stefan 2005).
4. DBP(FP) removal by AOPs
Disinfection byproduct formation potential (DBPFP) removal by AOPs
One of approaches to control DBPs in drinking water is to reduce DBP precursor, such as
NOM in the raw water by coagulation/flocculation or oxidation. The NOM removal in
coagulation/flocculation is quite low, between 10~50%. Therefore, AOP has been proposed as
an alternative for the control of DBP precursors (Wang, Hsieh et al. 2000; Chin and Berube
2005). AOP has been reported to reduce total organic carbon (TOC) and trihalomethane
formation potential (THMFP) of raw water by removing aromatic structures and double
bonds of NOM (Collivignarelli 2004). However little research has been performed on the
AOP effect on the haloacetic acid formation potential (HAAFP) (Chin and Berube 2005).
Ozone/UV was found to reduce THMFP and HAAFP by 80 and 70% respectively at an ozone
dose of 0.62 mgO3/mL and a UV dose of 1,610 mJ/cm2. Interestingly, Ozone had very little
impact on TOC concentration but rapidly reduced UV254 absorbance and reduced DBPFP,
which means ozone did not mineralize the NOM in the raw water but altered the chemical
structure of DBP precursors such that they did not form DBP (Chin and Berube 2005).
17
THMFP was found to decrease by either UV radiation or Vacuum UV(VUV) radiation, but
HAAFP decreased not by sole UV radiation but by VUV (Buchanan, Roddick et al. 2006).
DBP removal by AOPs
There have been only a few studies on DBPs removal by AOP, and reports are conflicting.
92~100% of 200 µg/L of chloroform, bromodichloromethane, dibromochloromethane and
bromoform were removed with 0.1% of H2O2 and 90 min of 3.2 mW/cm2 UV irradiation (UV
dose 17,280mJ/cm2) (Rudra, Thacker et al. 2005). However, in another study, THMs
increased at lower levels of UV/H2O2 doses and decreased with higher level of UV/H2O2
doses (Cassan, Mercier et al. 2006). In the other study, haloacetic acids (HAAs) decreased
with UV/H2O2 in two samples and increased in one sample (Paradis and Hoffman 2006). It
was reported that more bromine substituted THMs have slightly higher reaction rate constants
with hydroxyl radical (Mezyk, Helgeson et al. 2006). More bromine substituted THMs were
also shown to be better photolysed by UV (Nicole, De Laat et al. 1991).
DBP formation by AOPs
UV can produce similar DBPs to those formed by ozone or advanced oxidation process
(AOP) because UV radiation can result in the formation of ozone or radicals in water. UV
radiation was found to produce low levels of formaldehyde in surface water studies.
Formaldehyde concentration ranged up to 14 µg/L in UV treatment of raw water, whereas 1
to 2 µg/L levels were found in a UV treatment of conventionally treated water. However, the
overall effect of UV on DBPs was reported to be insignificant (EPA 1999).
The uniform formation conditions (UFC) test is the method to assess the effect of UV or
UV/H2O2 on DBP formation from subsequent chlorination or chloramination. H2O2 can react
with chlorine and DBP reagent affecting DBP formation and the measurement of chlorine
residuals due to higher oxidation potential than chlorine. Quenching H2O2 by bovine catalase
of 0.05-0.2 mg/L was proposed as a simple method that has no effect on DBP formation
(Muller and Jekel 2001; Liu, Andrews et al. 2003). In pilot or full scale plant, chlorine is used
to quench H2O2 (Royce and Stefan 2005).
Regarding DBP formation, UV/H2O2 has been reported to produce no harmful by-products
(Cotton and Collins 2006). Concentration of bromate (BrO3-), a DBP from the reaction of Br-
and ozone, from the ozonation coupled with UV was reported to be 40~50% lower than in
ozonation alone (Collivignarelli and Sorlini 2004).
18
5. Taste/odor and AOP
General concepts related with taste/odor
Drinking water taste and odor wheel
Suffet et al. updated the drinking water taste and odor wheel that consists of primary taste
and odor categories, common expressions of taste and odors from each categories, and typical
chemicals that cause the specific taste and odor (Suffet, Khiari et al. 1999) (Figure 2-1).
Sour
SweetSalty Bitter
Chemical
Mouth Feel
Earthy/ Musty
Chlorinous
Grassy/ Woody
SwampyFragrant
Fishy
Medicinal
Figure 2-1. The drinking water taste and odor wheel (Suffet, Khiari et al. 1999)
Sensory tests
Sensory tests evaluate the sensory characteristics of a sample and can be divided into two
categories: analytical and affective. Analytical tests use trained panelists and measure
characteristics of sample such as taste/odor attributes and intensity. Affective tests typically
use large numbers of untrained subjects and measure preference or acceptance to investigate
the consumer�s ability to detect a difference, reasons for detected difference, and attitudes
about the differences. Analytical tests can be divided into two categories: discriminative tests
and descriptive test. Discriminative tests determine if human perception is different between
samples. Generally, five panelists are recommended as a minimum number to reduce the
dominance by a single panelist. Triangle test, duo-trio test, and the 2-of-5 test are well known
19
discriminative tests. Descriptive tests are used for identifying the sensory characteristics,
correlating sensory test results to instrumental analysis. Four to fifteen trained panelists are
recommended. Attribute rating, flavor profile analysis (FPA), quantitative descriptive analysis
(QDA) are well known descriptive tests (Lawless and Heymann 1999; Meilgaard, Civille et
al. 1999).
Flavor profile analysis (FPA)
Five to eight panelists individually evaluate one sample at a time for both aroma and flavor
and record the attributes, aftertastes and intensities based on seven-point scale (none,
threshold, very slight, slight, slight-moderate, moderate, moderate-strong, strong). Discussion
among the panelists is allowed to reach a consensus on descriptors and intensity (Krasner,
McGuire et al. 1985; Meilgaard, Civille et al. 1999).
Weber-Fechner plot
Weber found that the amount of compounds added for the detectable change in intensity
increases in proportion to the initial concentration (Meilgaard, Civille et al. 1999).
Fechner derived an equation from the fact that plot of intensity perceived by panelists shows
logarithmic curve.
Weber-Fechner plot is a dose-response curve based on Weber-Fechner law shown below and
can be drawn from log concentration and taste/odor intensity (Rashash, Dietrich et al. 1997).
Where S is the average odor intensity, C is the concentration, a and b are the constants for
slope and intercept, respectively.
Earthy-musty odorants (Geosmin and 2-MIB)
Geosmin and 2-MIB are one of the most widespread odorants found in fresh water. Geosmin
and 2-MIB have been identified in fresh water as earthy-musty odorants, and reported to be
produced from algae or actinomycetes (Rashash, Dietrich et al. 1995; Suffet, Khiari et al.
1999; Jüttner and Watson 2007). These compounds cause seasonal earthy-musty odor episode
and are difficult to remove below threshold level by conventional water treatment due to the
20
poor removal efficiency and the low threshold level.
Other algae-related odorants
There is a �fishy/rancid� category in the drinking water taste and odor wheel. Fishy odors
were reported to occur naturally from the algae. 2-trans-4-cis-7-cis-decatrienal, trans-2, cis-4-
decadienal, n-heptenal, and trans,trans-2,4-heptadienal are typical fishy odorants in fresh
water, and 1-pentene-3-one was associated with rancid odors. Trans-2, cis-6-nonadienal,
cucumber-smelling aldehyde was reported to be produced from algae, and added in
�Fragrant: vegetable/fruity/flowery� category in the drinking water taste and odor wheel.
(Rashash, Dietrich et al. 1995; Suffet, Khiari et al. 1999; Watson, Satchwill et al. 2001).
Aldehydes were reported to play an important role in the production of off-flavor and have a
synergic effect with ketones or carboxylic acids (Andersson, Forsgren et al. 2005).
Oxidation of earthy-musty odorants (geosmin and 2-MIB)
Taste and odor episodes typically occur seasonally or periodically, mostly in warm summer
season, and it is difficult to predict when they occur and how long they last. Therefore,
sometimes it is not economical to install permanent treatment system such as granular
activated carbon (GAC) filter to control taste and odor. Especially for the utilities that use UV
for disinfection, adding H2O2 prior to UV step on an �as-needed� basis could be economic
and practical (Paradis and Hoffman 2006).
Glaze et al. investigated several types of AOPs as an alternative process for the removal of
2-MIB and geosmin. H2O2 or UV in addition to ozonation showed higher removal efficiency
(Glaze W. H. 1990). Complete removal of geosmin and 2-MIB was achieved with a
combination of 1.5~3 mg/L ozone (2~3 min contact time) and 500~600 mJ/cm2 UV radiation.
(Collivignarelli 2004). Addition of H2O2 in UV photolysis oxidized greater than 70% of 2-
MIB and geosmin while direct UV photolysis removed 10% and 25-50% of the 2-MIB and
geosmin at the UV dose of 1,000 mJ/cm2, respectively (Rosenfeldt 2005). In a pilot scale
study, optimal hydrogen peroxide dose of 6-10 mg/L was reported in terms of removal
efficiency, chlorine residual decay, and DBP formation (Paradis and Hoffman 2006).
Initially it was thought that UV could not perform both disinfection and advanced oxidation
in a system because of different levels of UV dose required. Recently, UV systems labeled as
�dual purpose� were developed and applied to full scale water treatment plants (WTPs).
21
These dual systems combine low intensity UV for disinfection and high intensity UV for both
disinfection and advanced oxidation of odorants (Cotton and Collins 2006). UV/H2O2 has
been applied to 9 full scale WTPs to control geosmin and MIB, N-nitrosodimethylamine
(NDMA), 1,4-dioxane, and PCE (Sarathy 2006). A pilot scale study in Canada reported that
site specific evaluation including impact on secondary disinfectant level and DBP formation
is required when the feasibility of UV/H2O2 on taste and odor control is investigated (Paradis
and Hoffman 2006).
Oxidation of algae-related odorants
Nonadienal had a greater reaction rate constant with hydroxyl radical than geosmin and 2-
MIB while other odorants such as 2-isopropyl-3-methoxypyrazine (IPMP), 2,4,6
trichloroanisole (TCA), and 2,6-di-tert-butyl-4-methylphenol (BHT) has similar or less
reaction rate constants compared to geosmin and 2-MIB (Peter and Von Gunten 2007). In
research on oxidation of algal metabolites, algal-related compounds were able to be degraded
by chlorine and permanganate. However, oxidation of certain algal metabolite caused the
formation of other odors (Dietrich, Hoen et al. 1995). Qualitative descriptors were reported to
change with odorant concentration change (Rashash, Dietrich et al. 1997).
Derivatization method for detecting carbonyl group
Carbonyls are frequently related with odors found in fresh water (Rashash, Dietrich et al.
1997; Suffet, Khiari et al. 1999; Watson, Satchwill et al. 2001; Satchwill, Watson et al. 2007),
and can be more easily determined by derivatization method. One method is the
derivatization with 2.4-dinitrophenylhydrazine (DNPH) followed by liquid-liquid extraction.
Another common method is the derivatization with pentafluorobenzyl-hydroxylamine
hydrochloride (PFBHA) followed by liquid-liquid extraction. Solid phase microextraction
(SPME) can be combined with these derivatization method for both liquid and headspace.
(Bao, Pantani et al. 1998).
22
6. Kinetics of geosmin/ 2-MIB and DBPs with hydroxyl radical
Second order rate constant of hydroxyl radical reaction in aqueous phase
Researchers have measured the second order reaction rate constants of odorants and DBPs
with hydroxyl radical as shown in Table 2-3. (Glaze, Schep et al. 1990; Mezyk, Helgeson et
al. 2006; Westerhoff, Nalinakumari et al. 2006; Cole, Cooper et al. 2007; Peter and Von
Gunten 2007). Reaction rate constants of geosmin, 2-MIB, and nonadienal are greater by
three orders of magnitude than those of THMs and chlorinated HAAs as shown. According to
the reaction rate constants, it is thought that DBPs can not be practically reduced by hydroxyl
radical reaction compared to odorants.
Table 2-3. Second order rate constants of DBPs and odorants with hydroxyl radical
Compounds Reaction rate constant with ·OH (M-1s-1)
Trichloromethane 0.7~5.4 x 107 a
Bromodichloromethane 7.1 x 107 a
Chlorodibromomethane 8.3 x 107 a THMs
Tribromomethane 1.5 x 108 a
Chloroacetic acid (MCAA) 8.3 x 107 b
4.0 x 108 c
4.3 x 107 d
Dichloroacetic acid (DCAA) 1.0 x 108 b HAAs
Trichloroacetic acid (TCAA) 6.0 x 107 b
1.4 x 1010 e Geosmin
7.8 x 109 f
8.2 x 109 e 2-MIB
5.1 x 109 f
Odorants
Nonadienal 10.5 x 109 f a Mezyk et al. 2006, b Maruthamuthu 1995, c Yokohata et al. 1969 d Adams et al. 1965 e Glaze et al. 1990, f Peter and Von Gunten 2007
23
7. Reaction mechanism of DBPs and Geosmin/2-MIB in UV/H2O2
Methods for investigating radical reaction mechanism
Electron Pulse Radiolysis (EPR) involves exposing γ-rays to an aqueous solution. The EPR
of water generates highly reactive electrons, radical ions, and neutral radical species
according to the following equation (Makogon, Fliount et al. 1998; Cole, Cooper et al. 2007).
The coefficients of species in the radiolysis are chemical yields, G which have a unit of
µmol/10J. γ-rays
H2O → [0.28]·OH + [0.06]H· + [0.27]eaq- + [0.05]H2 + [0.07]H2O2 + [0.27]H3O+
Laser flash photolysis, electron spin resonance (ESR) spectrometry, and spin trapping are
methods for investigating radical reaction mechanism. In the trapping method, a reactive
radical is trapped to form a more stable radical from which the structure of an initial radical
can be determined (Paul, Small et al. 1978; Kochany and Bolton 1992; Parsons 2000). For
organic pollutants, the reaction mechanism was investigated by analyzing intermediates and
final products using GC/MS, and by measuring total organic carbon (TOC) to make a carbon
balance in the process of the mineralization (Stefan, Hoy et al. 1996; Stefan and Bolton 1998;
Stefan, Mack et al. 2000). Bromide and chloride ion concentrations and pH change were
measured to investigate the mechanism in the reaction of halogenated compounds with
hydroxyl radical (Lay 1989; Milano, Bernatescallon et al. 1990; Crittenden, Hu et al. 1999;
Cole, Cooper et al. 2007) and in the reduction of haloacetic acid (Zhang, Arnold et al. 2004).
Geosmin/MIB oxidation by hydroxyl radical
No mechanism has been elucidated for the reaction of geosmin and 2-MIB with hydroxyl
radical. In a research of UV/H2O2, geosmin and 2-MIB was reported to be removed mainly
by hydroxyl radical reaction and partially by direct UV photolysis (Rosenfeldt, Melcher et al.
2005). Hydroxyl radical was reported not to be directly responsible for the degradation of
geosmin and 2-MIB in ultrasonication even though ultrasonication causes the oxidation by
hydroxyl radical as well as pyrolysis. In this research, degradation pathways were proposed
by identifying pyrolitic cyclo alkene intermediates using GC/MS analysis (Song and O'Shea
24
2007).
UV photolysis of Halogenated methanes
In regard to reaction mechanisms of DBP degradation by UV photolysis or advanced
oxidation process, only a few mechanisms of DBP degradation were studied. The UV
photolysis mechanism of tribromomethane (CHBr3) and carbon tetrabromide (CBr4) in
aqueous phase was proposed as a water-catalyzed dehalogenation. According to the proposed
mechanism, O-H was inserted and H-Br was eliminated by water-catalyzed reaction
producing three HBr and CO, and four HBr and CO2 as final products for tribromomethane
and carbon tetrabromide, respectively (Li, Kwok et al. 2004; Zhao, Lin et al. 2005). In these
studies, UV absorption spectra were measured from 190 to 280 nm wavelength to make a
mass balance of bromide and hydrogen ion. Bromide concentration was measured from the
increased UV absorbance at 190nm. Decreased concentration of tribromomethane was
measured from the decreased UV absorbance at 215nm. Final products were detected by 13C
NMR, infrared spectrum and Raman shift. The reaction pathway was proposed as below.
Reaction mechanisms of other THMs have not been elucidated.
CHBr3 + hν → ·CHBr2 + ·Br
·CHBr2 + ·Br → BrCHBr-Br (isobromoform)
BrCHBr-Br + n(H2O) → CHBr2OH + HBr + (n-1) H2O
CHBr2OH + n(H2O) → HBrCO + HBr + n(H2O)
HBrCO + n(H2O) → CO + HBr + n(H2O)
---------------------------------------------------------------------
Overall CHBr3 + hν +n(H2O) → CO + 3HBr + (n-1)H2O
Hydroxyl radical reaction of halogenated methane
The first step of the reaction mechanism of tribromomethane and hydroxyl radical in the
gas phase was proposed to be hydrogen abstraction. The ·CBr3 radical produced from the
reaction was proposed to be degraded in two pathways, which are reaction with hydroxyl
radical, and more likely, with oxygen (Fliount, Makogon et al. 1997; McGivern, Francisco et
al. 2002; McGivern, Kim et al. 2004).
In gas phase :
CHBr3 + ·OH → ·CBr3 + H2O
Pathway Ⅰ
25
·CBr3 + ·OH → CBr3OH
CBr3OH + H2O → 3H+ + 3Br- + CO2
Pathway Ⅱ
·CBr3 + O2 → CBr3OO· → → → 3 Br- + CO2 (preferred)
Water catalysis was reported in the hydroxyl radical reaction of acetaldehyde where the
reaction rate was enhanced in the presence of water. This increased reaction rate was
explained by the reduction of an intrinsic reaction barrier resulting from the water
aggregation (Vohringer-Martinez, Hansmann et al. 2007).
Polar effects and deuterium isotope effect in hydroxyl radical reaction
Reactivity of an atom or radical had a more direct relationship with stabilization of the
transition states by the polar effect rather than exothermicity. The factor of transition state
energy difference to enthalpy change between normal and deuterated reactants (α) could be
used as a measure of reactivity of an atom or radical, and could be interpreted as a percentage
of C-H bond breakage (Russell 1957). Kinetic-isotope effect(KIE) is the ratio of reaction rate
constant between original compound and deuterated compound as presented below, and can
be used to elucidate the reaction mechanism such as hydrogen abstraction (Russell 1957;
Farkas, Szilagyi et al. 2003). KIE varies with the types of compounds ranging from 1.0 to
11.9, and KIE of 5.7 was reported for the hydroxyl radical reaction of acetone where
hydrogen abstraction was attributed to 50% reaction. However, KIE can give the information
only on the rate-controlling step of the reaction, and in itself, is not sufficient for elucidating
the reaction mechanism because the contribution of hydrogen abstraction compared to other
pathway also has to be known (Farkas, Szilagyi et al. 2003).
kH
CClH2COOH + ·OH → ·CClHCOOH + H2O
→ other products
kD
CClD2COOD + ·OH → ·CClDCOOD + HDO
→ other products
kH /kD = Deuterium isotope effect
26
Peroxy radical reaction
In the presence of oxygen, radical species produced from carbon-halogen bond cleavage,
hydrogen abstraction, and electron transfer reaction react with oxygen and form the peroxyl
radical (Spangenberg, M?ler et al. 1996; Lifongo, Bowden et al. 2004; Zalazar, Labas et al.
2007). In the Russell mechanism, peroxyl radicals mutually react with each other and
generate a ketone and a alcohol, while two ketones and hydrogen peroxide are produced in
the Bennett mechanism (Russell 1957; Bennett and Howard 1973).
Halogenated acetic acid
For the hydroxyl radical reaction of trichloroacetic acid (TCAA), photo-Kolbe reaction was
proposed as a reasonable mechanism because hydrogen abstraction was impossible. This
Kolbe reaction was thought to be more effective for less halogenated HAAs due to higher
electron density at the carboxyl function (Mao, Schoeneich et al. 1991).
CCl2HCOO· → ·CHCl2 + CO2 (Kolbe mechanism)
In research on radical-meditated degradation of tribromoacetic acid (TBAA), hydroxyl
radical was likely to indirectly oxidize TBAA by oxidation of bromide (Fliount, Makogon et
al. 1997).
·OH + 2Br- → Br2·- + OH-
Br2·- ↔ Br· + Br-
Br· + CBr3COO- → Br- + CBr3CO2·
CBr3CO2· → ·CBr3 + CO2
·CBr3 + O2 → CBr3OO· →→→ 3Br- + CO2
Hydrogen abstraction and electron transfer reaction were proposed to be the first step in the
degradation mechanism of dichloroacetic acid (DCAA) by UV/H2O2, and HCl and CO2 were
proposed as final products (Zalazar, Labas et al. 2007). In this study, chloride and total
organic carbon concentration were measured and plotted with calculated values. Two moles
of chloride and hydrogen ion were shown to be produced from each mole of DCAA, and
complete mineralization was achieved based on the molar decrease in TOC. Reaction
27
mechanism was proposed as follows:
Pathway Ⅰ
CCl2HCOO- + ·OH → ·CCl2COO- + H2O
·CCl2COO- + O2 → ·OOCCl2COO-
·OOCCl2COO- → COCl2 + CO2 + 1/2O2
COCl2 + H2O → CO2 + 2HCl
Pathway Ⅱ
CCl2HCOO- + ·OH → CCl2HCOO· + HO-
CCl2HCOO· → ·CHCl2 + CO2 (Kolbe mechanism)
·CHCl2 + O2 → Cl2HCOO·
Cl2HCOO· → COCl2 + 1/2H2O2
COCl2 + H2O → CO2 + 2HCl
In research on TCAA degradation in the gas phase, photochemical disproportionation was
proposed as below (Spangenberg, M?ler et al. 1996):
CCl3COOH + hν + H2O + 1/2 O2→ 3HCl + 2CO2
CCl3COOH + hν + H2O → 3HCl + CO + CO2 (In acidic solution)
A trace amount of trichloromethane was observed from the degradation of TCAA in the
same research, which was explained by the following reaction:
CCl3COOH → CHCl3 + CO2
Chemiseddine and Boehm (1990) obtained a Cl-/CO2 ratio of about 2:1 from the
photocatalytical degradation of TCAA. In this research, the slow reaction rate of TCAA to
MCAA was explained by the absence of α-C-H bond where hydrogen could be abstracted by
hydroxyl radical. Anglada (2004) reported that hydroxyl radical predominantly extracted
acidic hydrogen of formic acid by electron transfer mechanism while hydrogen abstraction
from carbon contributed at higher temperatures.
In the case of photodegradation of HAAs, the rates were proportional to the number of
halogen atoms (i.e. TCAA>DCAA>MCAA). The final products from the photodegradation
of HAAs were HCl and CO2, which indicates complete mineralization. The main process for
the photodegradation of HAAs is proposed as the C-X bond cleavage where electronegativity
28
of halogen atoms plays an important role to the bond strength. However, at higher
temperatures, reaction rate increased because of thermal decarboxylation. Apparent reaction
rate of photocatalytic dehalogenation of HAAs were in the order TCAA > TBAA > DCAA >
DBAA > MCAA > MBAA (Lifongo, Bowden et al. 2004).
Li et al. proposed decarboxylation for the hydroxyl radical reaction in the photocatalytic
degradation of MCAA and DCAA, while TCAA was shown to have a different mechanism
(Li, Xie et al. 2006).
TiO2 + hν → ecb- + hvb
+
hvb+ + OH- → ·OH
MCAA
CH2ClCOO- + ·OH→ CO2 + ·CH2Cl + OH-
·CH2Cl + H2O → CH2Cl-OH + H·
CH2Cl-OH + H2O → CH2(OH) 2 +HCl
DCAA
CHCl2COO- + ·OH→ CO2 + ·CHCl2 + OH-
·CHCl2 + H2O → CHCl2-OH + H·
In the sonolysis of TCAA, two mechanisms were proposed; free radical reaction and thermal
degradation (Wu, Wei et al. 2001). ))), ·OH
CCl3COO- → CCl3COO· → → Cl- + CO2 + CO + H2O ))), ∆
CCl3COO- → other intermediates → → Cl- + CO2 + CO + H2O
Halogenated organic compounds
Oxidation of 1,1,1-trichloroethane and halomethane with hydroxyl radical was proposed to
be initiated by hydrogen abstraction (Makogon, Fliount et al. 1998; Louis, Gonzalez et al.
2000; Louis, Gonzalez et al. 2000; Louis, Gonzalez et al. 2001) and hydrogen-abstracted-
radical was proposed to react with oxygen producing another intermediate peroxyl radical
(Makogon, Fliount et al. 1998).
In aqueous phase :
29
CHXYZ + ·OH ! ·CXYZ + H2O (X, Y, Z = H, Cl, Br, or F)
CCl3CH3 + ·OH ! CCl3·CH2 + H2O
CCl3·CH2 + O2 ! CCl3CH2OO·
The primary mechanism for the reaction of acetic acid and hydroxyl radical in the
atmosphere was suggested as the abstraction of the acidic (carboxyl group) hydrogen
(Butkovskaya, Kukui et al. 2004; Vimal and Stevens 2006) and reaction pathway was
postulated as follows;
CH3COOH + ·OH → CH3COO· + H2O → ·CH3 + CO2 + H2O
Chemical properties of C-H bond in halogenated DBPs
According to the studies on the gas phase reaction of hydroxyl radical and halogenated
methane and acid, both chlorinated and brominated compounds have similar activation
energies and reaction enthalpy changes (Louis, Gonzalez et al. 2000; Lagoa, Diogo et al.
2001). In addition, in the transition state of the gas phase reaction between halogenated
methane and hydroxyl radical, C-H bond lengths, bond angles and ratios between the
elongation of the C-H bond and O-H bond of brominated and chlorinated methane are very
close to each other (Louis, Gonzalez et al. 2000). C-H and C-Cl bond lengths in transition
state were shown to slightly increase with the increasing number of chlorine atoms, and
reaction of trichloromethane and hydroxyl radical was more favorable than
monochloromethane (Louis, Gonzalez et al. 2000; Louis, Gonzalez et al. 2004). C-Cl bond is
stronger than C-Br bond in halomethanes (80.1 and 70.4 kcal/mol respectively) and in
haloacetic acid (McGivern, Derecskei-Kovacs et al. 2000; McGivern, Derecskei-Kovacs et al.
2000).
30
Chapter 3. Simultaneous Removal of DBPs and Odorants
by UV/H2O2 Advanced Oxidation Process
Submitted to Water Research (June 2008)
Abstract Many utilities experience both taste/odor episodes and higher disinfection byproduct level
mostly in summer. This research investigated if UV/H2O2, when applied for the removal of
odorants geosmin and 2-methyl isoborneol, could simultaneously remove trihalomethane and
haloacetic acid disinfection byproducts. These results demonstrate that brominated
trihalomethanes and haloacetic acid were substantially removed by direct UV photolysis in
UV/H2O2 at the same dose for removing geosmin and 2-MIB. Tribromomethane and
dibromochloromethane were removed by 99% and 80% respectively at the UV dose of 1,200
mJ/cm2 and 6 mg/L H2O2, where geosmin and 2-MIB were removed by 95% and 65%
respectively. Tribromoacetic acid (TBAA) and dibromoacetic acid (DBAA) were removed by
99% and 90% respectively under the same condition. Brominated DBPs were removed by
direct photolysis, presumably via photo-induced C-Br bond cleavage. Concentrations of
trichloromethane and chlorinated HAAs were not substantially reduced under the same
conditions. Reduction of brominated DBPs can be a significant addition to water utilities that
have difficulty in meeting regulated DBPs level especially in the region with higher bromine
concentration. These results indicate that the UV/H2O2 can be seasonally applied to control
both taste/odor and brominated DBPs.
31
Introduction
Advanced oxidation process (AOP) in water treatment involves the hydroxyl radical (�OH).
AOP essentially mimics photo-initiated oxidation processes in natural systems, such as sun
light on surface water or in the atmosphere (Oppenlander 2003). AOP has been proven to
efficiently remove organic contaminants without production of residual solids, which is an
advantage compared to the activated carbon adsorption process. Ultraviolet (UV) irradiation
is well established for disinfection of water. UV/H2O2 process is a homogeneous AOP in
which hydroxyl radicals are generated by the direct photolysis of H2O2 under UV irradiation
(Liao and Gurol 1995; Stefan, Hoy et al. 1996; Stefan and Bolton 1998; Stefan, Mack et al.
2000; Rosenfeldt, Melcher et al. 2005; Rudra, Thacker et al. 2005; Xu, Gao et al. 2007). This
process results in highly efficient removal of organic contaminants, including recalcitrant
odorous compounds such as geosmin and 2-MIB, mainly by the reaction with hydroxyl
radicals and partially by direct UV photolysis (Beltran, Ovejero et al. 1993; Stefan, Hoy et al.
1996; Stefan and Bolton 1998; Cater, Stefan et al. 2000; Stefan, Mack et al. 2000; Rosenfeldt,
Melcher et al. 2005; Rudra, Thacker et al. 2005; Paradis and Hoffman 2006; Rosenfeldt and
Linden 2007). Recently, UV systems labeled as �dual purpose� were developed and applied
to full scale water treatment plants (WTPs). These dual systems combine low intensity UV
for disinfection and high intensity UV for both disinfection and advanced oxidation of
odorants (Cotton and Collins 2006). The operation of these systems involves UV
transmittance (UVT) and alkalinity constraints because these increase the demand for
hydroxyl radical (Ho, Croue et al. 2004; Cotton and Collins 2006). The optimum H2O2 dose
in the UV/H2O2 process should be empirically determined because excess H2O2 can be an
hydroxyl radical scavenger (Wang, Hsieh et al. 2000).
Disinfection byproducts (DBPs) and taste/odor compounds are two of the major problems in
drinking water quality. DBPs form from the reaction of DBP precursors and disinfectant.
Natural organic matter (NOM) such as humic or fulvic acid, is a typical DBP precursor. AOP
has been proposed as an alternative method to reduce disinfection byproduct formation
potential (DBPFP) by reducing total organic carbon (TOC), and aromatic structures or double
bonds of NOM (Kusakabe, Aso et al. 1990; Wang, Hsieh et al. 2000; Murray and Parsons
2004). Trihalomethanes (THMs) and haloacetic acids (HAAs) represent regulated DBPs in
drinking water. Chlorinated compounds such as trichloromethane or trichloroacetic acid are
the most prevalent DBPs. However, THMs can be locally composed of more than 40% of
32
brominated THMs, and HAAs can be composed of 10-25% of brominated HAAs (Hyun, Kim
et al. 2005; Buchanan, Roddick et al. 2006). Brominated HAAs were reported to constitute at
least 10% of the total HAA concentration in waters containing 0.1mg/L bromide (Cowman
and Singer 1996). In addition, brominated DBPs are more toxic than their chlorinated
analogues (Echigo, Itoh et al. 2004; Richardson, Plewa et al. 2007) and can be problematic in
regions where aqueous bromide concentrations are relatively higher.
Geosmin and 2-MIB are typical earthy-musty odor compounds found in surface water and
drinking water resulting in seasonal odor episodes, and mostly related to cyanobacteria or
actinomycetes (Jüttner and Watson 2007). These compounds are difficult to remove by
conventional water treatment processes and have low odor threshold levels (4-10 ng/L); thus
activated carbon or AOPs are required to control them.
Advanced technologies are expected to control multiple contaminants in full scale WTPs. If
an advanced process can reduce both DBPs and odorous compounds significantly, it would
be immensely beneficial to many WTPs. Many studies reported AOPs can control DBP
precursors and consequently reduce DBP level in finished water (Wang, Hsieh et al. 2000;
Chin and Berube 2005; Buchanan, Roddick et al. 2006). However, many WTPs are utilizing
pre-chlorination to control taste/odor or iron/manganese or ammonia nitrogen or to obtain
required CT value, which occurs prior to coagulation and produces a variety of DBPs while
AOPs typically occur after filtration to increase UV transmission. Therefore, in case of pre-
chlorination, DBPs are already formed before filtration. There have been only a few studies
on the removal of DBPs by UV/H2O2, and these reports are contradictory. Rudra et al. (2005)
reported over 90% removal of THMs at high UV and H2O2 dose (17,000 mJ/cm2 and 0.1%
respectively). In another study, THMs increased at lower levels of UV/H2O2 doses and
decreased with higher level of UV/H2O2 doses, and HAAs decreased for two samples and
increased for one sample (Paradis and Hoffman 2006).
Researchers have measured second order rate constants for the reaction of odorous
compounds and DBPs with hydroxyl radical (Mezyk, Helgeson et al. 2006; Westerhoff,
Nalinakumari et al. 2006; Cole, Cooper et al. 2007; Peter and Von Gunten 2007). owever, few
studies have been reported on the possibility and mechanisms of DBPs removal by UV/H2O2
process in aqueous phase.
There are two mechanisms associated with UV/H2O2 oxidation treatment. One involves
photolysis of H2O2, yielding the hydroxyl radical which subsequently reacts with the
contaminant, generally by abstracting a hydrogen, or by adding to an unsaturated site. The
33
other mechanism involves direct photolysis of the contaminant itself, often resulting in bond
homolysis and radical generation. These radicals subsequently are oxidized by reaction with
H2O2, O2, etc. Direct UV photolysis has shown to be either a partial or substantial contributor
for the removal of organic compounds based on the types of compounds when UV/H2O2 is
applied. Geosmin and 2-MIB decreased by 40% and 20% respectively with a UV dose of
1,700 mJ/cm2 (Rosenfeldt, Melcher et al. 2005), diazinon decreased by 20% at a UV dose of
600 mJ/cm2 (Shemer and Linden 2006), and microcystin decreased by 50% at a UV dose of
approximate 3,000 mJ/cm2 by direct UV photolysis (Qiao, Li et al. 2005). In regard to UV
photolysis of THMs, it was reported that only the brominated THMs were photolysed and
quantum yield of the photolysis was 0.43 (Nicole, De Laat et al. 1991). In the same research,
polybrominated THMs were shown to be photolysed faster. In other research,
tribromomethane and chlorodibromomethane levels in chlorinated swimming pool water
were reported to decrease significantly with UV irradiation of 145 mJ/cm2 (Cassan, Mercier
et al. 2006). Polyhalomethanes such as tribromomethane (CHBr3) and carbon tetrabromide
(CBr4) were reported to be photolysed by a proposed water-catalyzed O-H insertion/HBr
elimination (Li, Kwok et al. 2004; Zhao, Lin et al. 2005).
For the direct UV photolysis and hydroxyl radical reactions of organic compounds such as
geosmin/2-MIB, diethyl phthalate, and pharmaceutical compounds in UV/H2O2, pseudo-first
order reaction models at a wavelength (λ) were proposed as follows (Sharpless and Linden
2003; Rosenfeldt, Melcher et al. 2005; Pereira, Weinberg et al. 2007; Xu, Gao et al. 2007):
[ ] '[ ]d C k Cdt
− =
Where, ' ' 'd ik k k= +
'k = the observed pseudo-first order rate constant (s-1)
' dk = the measured pseudo-first order rate constant of direct photolysis (s-1)
= ( )Sk λ ( )c λΦ ( )C λΦ = quantum yield for the photolysis of compound (mol Es-1)
( )Sk λ = specific rate of UV absorption by compound (Es mol -1s-1)
= 0 ( )( ) ( )[1 10 ]
( )
a zpE
a z
λλ ε λλ
−−
0 ( )pE λ = incident photon irradiance (mEscm-2s-1)
( )ε λ = molar extinction coefficient of compound at a specific wavelength (M-1cm-1)
34
( )a λ = solution absorbance at a specific wavelength (cm-1)
z = solution depth (cm)
' ik = the measured pseudo-first order rate constant of the reaction with ·OH
/' [ ]i C OH ssk k OH=
/C OHk = Second order reaction rate constant of compound and ·OH (M-1s-1)
[ ]ssOH = steady state ·OH concentration (M)
In case of the hydroxyl radical reaction, a steady state radical concentration is assumed due
to relatively higher H2O2 concentration (0-30 mg/L) (Sharpless and Linden 2003; Rosenfeldt,
Melcher et al. 2005; Pereira, Weinberg et al. 2007; Xu, Gao et al. 2007).
2 2, 2 2
,
( ) ( )[ ][ ]
[ ]S H O OH
ssS OH i
i
k H OOH
k Sλ λΦ
=∑∑
2 2, ( )S H Ok λ = specific rate of UV absorption by H2O2 (Es mol -1s-1)
( )OH λΦ = quantum yield for ·OH formation (≈ 1mol Es-1)
,S OHk
= second order reaction rate constant of scavenging species and ·OH (M-1s-1)
[ ]iS
= concentration of scavenging species (M)
Second order reaction rate constants of hydroxyl radical with geosmin and 2-MIB are 0.78 x
109 ~ 1.4 x 1010 M-1s-1, which are greater by three orders of magnitude than for the reaction of
hydroxyl radical with THMs (0.7 x 107 ~1.5 x 108 M-1s-1) or chlorinated HAAs (6 x 107 ~1.0
x 108 M-1s-1). The reaction rate constant of tribromomethane is greater than trichloromethane
by a factor of 10 (Glaze, Schep et al. 1990; Maruthamuthu, Padmaja et al. 1995; Mezyk,
Helgeson et al. 2006; Westerhoff, Nalinakumari et al. 2006; Cole, Cooper et al. 2007; Peter
and Von Gunten 2007).
This research investigated simultaneous removal of odorants and DBPs under conditions
similar to when UV/H2O2 is applied for removing recalcitrant odorants. The objectives of this
research were to investigate the types of DBPs that can be removed while exposed to
UV/H2O2 doses designed for geosmin/2-MIB control at typical concentrations found in
drinking water, and to evaluate the role of UV photolysis and hydroxyl radical reaction
involved in this removal.
35
Methods and Materials
1. Apparatus Experiments were performed with a Rayonet RPR-100 photochemical reactor equipped with
253.7 nm wavelength UV lamps of 7.2 mW/cm2 total intensity, and quartz reaction vessels.
UV dose was confirmed with the iodide/iodate actinometer (Rahn 2004; Rahn, Bolton et al.
2006). Samples were completely mixed and headspace free while being irradiated with UV
(Figure 3-1).
Figure 3-1. UV irradiation system and quartz reactor
2. Reagents and Sample Preparation Samples were prepared in de-ionized water using individual compounds; geosmin (200
mg/L, Supelco), and 2-MIB (100 mg/L, Supelco), trichloromethane (≥99%, Fisher scientific),
tribromomethane (≥99%, Acros organics), chloroacetic acid (MCAA) (≥99%, Aldrich),
dichloroacetic acid (DCAA) (≥99%, Signa-Aldrich), trichloroacetic acid (TCAA) (≥99%,
Alfa Aesar), bromoacetic acid (MBAA) (≥99%, Sigma-Aldrich, dibromoacetic acid (DBAA)
(≥99%, Fluka), tribromoacetic acid (TBAA) (≥99%, Acros organics). Hydrogen peroxide
(30%, Fisher) was diluted to desired concentrations of 6 mg/L which was selected based on
typical concentration range in pilot scale study (Paradis and Hoffman 2006), and added into
the samples immediately before UV irradiation. THMs standard (5,000 mg/L, Ultra
Scientific) was used in THM mixture samples for comparing the removal rates in UV/H2O2.
HAA9 standard mixture could not used because it was dissolved in tert-methyl butyl ether
(MTBE) that has a great scavenging effect with hydroxyl radical (second order rate constant,
36
k=3.9 x 109 M-1s-1) (Chang and Young 2000). Typical concentrations of compounds used in
the research are shown in Table 3-1.
Table 3-1. Typical concentrations of compounds in the research
Typical concentrations in the research Compounds
µg/L µM Odorants
geosmin
2-MIB
0.04-0.2
0.1-0.3
0.0002-0.0005
0.0006-0.002
THMs
Trichloromethane
Bromodichloromethane
Dibromochloromethane
Tribromomethane
60-500
90
80
80-550
0.5-4.2
0.5
0.4
0.3-2.2
Tetrahalo methanes
Carbon tetrachloride
Carbon tetrabromide
350
1,000
2.3
3.0
HAAs
Chloroacetic acid (MCAA)
Dichloroacetic acid (DCAA)
Trichloroacetic acid (TCAA)
Bromoacetic acid (MBAA)
Dibromoacetic acid (DBAA)
Tribromoacetic acid (TBAA)
270
190
180
200
190
160
2.9
1.5
1.1
1.4
0.9
0.5
Hydrogen Peroxide 6,000 176.5
3. Analysis Geosmin and 2-MIB were measured by solid-phase microextraction (SPME, Supelco) with
GC/MS (Agilent 5973) as in other studies (Watson, Brownlee et al. 1999; Watson, Brownlee
et al. 2000; Song and O'Shea 2007). Compounds partitioned from sample water were sorbed
on SPME fiber (65µm, PDMS/DVB) for 10 min at 60℃. The SPME fiber was injected into
the GC at 220℃ and desorbed for 2 min. A Rtx-5Sil column (30m, 0.25mm ID) with a
temperature program of 60℃ to 180℃ by 15℃/min was used. Approximate retention times
of 2-MIB and geosmin were 5.4 min and 7.9 min respectively; m/z value of 112, 125, 182 for
geosmin and 95, 108, 168 for 2-MIB were detected in selective ion mode. THMs were
37
measured based on Standard Method 6232.D by purge/trap (Tekmar 3000) and GC
(Tremetrics 9001) with DB-624 column (J & W). GC temperature was initially maintained at
45℃ for 3min, and then increased by 11℃/min up to 200℃. HAAs were determined by
liquid-liquid extraction method (EPA 552.2 method) and GC (HP 5890) ECD detector.
Injector temperature was 210℃ and initial oven temperature was set to 35℃ and increased up
to 140℃. UV absorption was measured by UV/Vis spectrophotometer (Beckman, DU640).
H2O2 concentration was determined by triiodide (I3-) titration method (Klassen, Marchington
et al. 1994). Linear regression was performed to determine the difference between reaction
rates of compounds under UV photolysis and hydroxyl radical reaction (α=0.05).
Results
1. UV absorbance Direct UV photolysis has been known to partially reduce organic compounds in UV/H2O2
process, although hydroxyl radicals are thought to play the key role (Liao and Gurol 1995;
Stefan, Hoy et al. 1996; Stefan and Bolton 1998; Stefan, Mack et al. 2000; Sharpless and
Linden 2003; Rosenfeldt, Melcher et al. 2005; Rudra, Thacker et al. 2005; Pereira, Weinberg
et al. 2007; Peter and Von Gunten 2007; Xu, Gao et al. 2007). To investigate the relative role
of hydroxyl radical production versus UV photolysis of the organic contaminants in UV/H2O2
process, molar extinction coefficients were measured (Figure 3-2). Brominated compounds
and geosmin and 2-MIB had at least two order of magnitude higher molar extinction
coefficients than chlorinated compounds, and one order of magnitude than H2O2. At typical
concentrations used in this research, H2O2 of mg/L concentration and brominated DBPs of
µg/L concentration were shown to absorb UV appreciably. Geosmin and 2-MIB of ng/L
concentration barely absorb UV. UV absorbance result suggests that one mechanism via
which brominated DBPs can be reduced is by UV photolysis. H2O2 of 6 mg/L absorbed the
UV the most at the wavelength of 253.7 nm, which indicates that hydroxyl radicals can be
produced from UV photolysis of H2O2 under these conditions.
38
Figure 3-2. Molar absorption coefficients measured at 254 nm in this research
2. Removal of odorants and DBPs at typical concentrations found in
drinking water
Geosmin/2-MIB and DBPs were prepared in a mixed sample. The reaction rates were
compared under the same UV/H2O2 condition to investigate types of DBPs whose
concentrations were reduced at the UV/H2O2 dose effective for removing geosmin/2-MIB.
These results demonstrate that for both THMs and HAAs, brominated DBPs concentrations
were not only reduced faster than chlorinated DBPs, but they could be completely or partially
removed at the UV/ H2O2 dose for removing geosmin/2-MIB.
2.1 Geosmin and 2-MIB
Geosmin and 2-MIB results showed 90 and 65 % removal, respectively, with a UV dose of
1,200 mJ/cm2 and 6 mg/L H2O2. Under identical conditions, but in the absence of H2O2, only
about 20% were removed with UV photolysis (Figure 3-3). As suggested by other research
(Rosenfeldt, Melcher et al. 2005), geosmin and 2-MIB concentrations are mainly reduced by
reaction with hydroxyl radical (formed by photolysis of H2O2) rather than by direct
photolysis of these compounds.
39
Figure 3-3. Comparison of removal rate between UV photolysis and UV/H2O2 for geosmin and
2-MIB
Initial concentration (C0) : geosmin (no H2O2) = 39.9 ng/L, geosmin (H2O2 6mg/L) = 183.4 ng/L, 2-MIB (no
H2O2) = 108.0 ng/L, 2-MIB (6mg/L H2O2) = 306.4 ng/L
2.2 Geosmin/MIB and THMs
THMs removal compared to geosmin/2-MIB by UV/H2O2
Brominated THMs were shown to be simultaneously removed at the UV/H2O2 dose for
removing geosmin/2-MIB. Tribromomethane and dibromochloromethane were removed by
99% and 80% respectively at the UV dose of 1,200 mJ/cm2 and 6 mg/L H2O2, where geosmin
and 2-MIB were removed by 95% and 65% respectively. The THMs with higher numbers of
bromine atoms were removed faster than trichloromethane, which for all practical purposes,
was not removed by UV/ H2O2 (Figure 3-4). Tribromomethane was removed faster than
either geosmin/2-MIB or other THMs.
40
Figure 3-4. Removal of geosmin/2-MIB and THMs with UV/H2O2
Initial concentration (C0) : geosmin = 43.3 ng/L, 2-MIB = 100.0 ng/L, trichloromethane = 63.2 µg/L,
bromodichloromethane = 93.5 µg/L, dibromochloromethane = 75.7 µg/L, tribromomethane = 81.9 µg/L, H2O2 =
6 mg/L
Direct UV photolysis of brominated THMs
To investigate the contribution of direct UV photolysis on the removal of compounds, a
mixture of THMs was reacted with UV in the absence of H2O2. Brominated THMs were
removed by direct UV photolysis and removal rates were in direct proportion with the
number of bromine atoms in their molecule. For three brominated THMs, there was no
significant difference between the removal rates of UV photolysis and UV/H2O2 treatment as
a result of linear regression analysis (α=0.05, for bromodichloro methane p=0.56,
dibromochloromethane p=0.17, tribromomethane p=0.51) (Figure 3-5) indicating that
brominated THMs are removed, not by reaction with hydroxyl radical, but rather by direct
UV photolysis, with C-Br bond cleavage as the likely first step of the process (Li, Kwok et al.
2004; Zhao, Lin et al. 2005).
41
Figure 3-5. Comparison of removal rates between UV photolysis and UV/H2O2 for brominated
THMs Initial concentration (C0) : trichloromethane = 63.2 µg/L, bromodichloromethane = 93.5 µg/L,
dibromochloromethane = 75.7 µg/L, tribromomethane = 81.9 µg/L, H2O2 = 6 mg/L
Removal mechanism of THMs
To investigate the role of hydrogen abstraction and its effects on THMs removal, reaction of
carbon tetrachloride (CCl4) and carbon tetrabromide (CBr4), were compared to each other,
and to trichloromethane or tribromomethane respectively. Because they do not possess
abstractable hydrogens, carbon tetrachloride and carbon tetrabromide cannot react with
hydroxyl radical and can only be removed by direct photolysis. Based on these result, the tri
and tetra brominated methanes were removed faster than their chlorinated analogues as
shown in Figure 3-6. Furthermore, CX4 (X = Cl or Br) were removed at a greater rate than
CHX3, even though the latter posesses a hydrogen atom that can be abstracted by hydroxyl
radical. These results confirm that the different removal rates in UV/H2O2 between
chlorinated and brominated THMs results from the different UV photolysis rates and not
from the hydrogen abstraction by hydroxyl radicals.
42
Figure 3-6. Removal rates of halogenated methanes measured for individual compounds
Initial concentration (C0) : trichloromethane = 513.6 µg/L, tribromomethane = 523.3 µg/L, carbon tetrachloride
= 343.3 µg/L, carbon tetrabromide 926.3 µg/L, H2O2 = 6 mg/L
2.3 Geosmin/2-MIB and HAAs
HAAs removal compared to geosmin/2-MIB by UV/H2O2
Treatment with UV/H2O2 removed brominated HAAs faster than chlorinated HAAs (Figure
3-7). Tribromoacetic acid (TBAA) and dibromoacetic acid (DBAA) were removed by 99%
and 90% respectively at the UV dose of 1,200 mJ/cm2 and 6 mg/L H2O2, where geosmin and
2-MIB were removed by 95% and 65% respectively. Chlorinated HAAs with no bromine
atoms were barely removed by a UV dose range 0 ~ 4,300 mJ/cm2 and 6 mg/L of H2O2 that
would effectively remove geosmin and 2-MIB. Brominated HAAs removal rates increased in
proportion to the number of bromine atoms in the molecule. Consequently, tribromoacetic
acid (TBAA) had the highest removal rate among all HAAs and geosmin/2-MIB. During
water treatment, TBAA and DBAA can be substantially removed at the UV dose for
removing geosmin/2-MIB in UV/H2O2 process while MBAA was not efficiently reduced.
43
Figure 3-7. Removal of Geosmin/2-MIB and HAAs with UV/H2O2 Initial concentration (C0) : geosmin = 183.4 ng/L, 2-MIB = 306.4 ng/L, bromoacetic acid = 202.4 µg/L,
dibromoacetic acid = 190.4 µg/L, tribromoacetic acid = 161.2 µg/L, chloroacetic acid = 270.7 µg/L,
dichloroacetic acid = 190.6 µg/L, trichloroacetic acid = 175.9 µg/L, H2O2 = 6 mg/L
UV photolysis of brominated HAAs
Three brominated HAAs were reacted with only UV to investigate the contribution of UV
photolysis on the higher removal of brominated HAAs in UV/H2O2 process. As shown in
Figure 3-8, no significant difference of removal rate were observed for TBAA between UV
photolysis and UV/H2O2 process (α=0.05, p=0.3), and slightly better removals were observed
with UV photolysis for MBAA and DBAA (α=0.05, p<0.002) from the linear regression
analysis. Thus, brominated HAAs are removed mainly by UV photolysis in UV/H2O2 process.
44
Figure 3-8. Comparison of removal rates between UV photolysis and UV/H2O2 for brominated
HAAs
Initial concentration (C0) : bromoacetic acid = 202.4 µg/L, dibromoacetic acid = 190.4 µg/L, tribromoacetic acid
= 161.2 µg/L, H2O2 = 6 mg/L
2.4 Removal efficiency of DBPs in the presence of inorganic ions
In order to investigate the removal efficiency of DBPs in the presence of inorganic ions,
percent removals of selected DBPs in reference water which contains 50mg/L of alkalinity
were compared. After 5 min UV irradiation, relative % removal of MCAA,
dibromochloromethane (DBCM), and chlorodibromoacetic acid (CDBAA) in reference water
were shown to be similar to de-ionized water (Table 3-2). This result indicates that relative
removal efficiency of DBPs compared to geosmin and 2-MIB would not change with the
different water matrix.
Table 3-2. Comparison of % removal in de-ionized water and reference water
Relative % removal to 2-MIB Compound
De-ionized water Reference water
Chloroaectic acid (MCAA) 31 28 Dibromochloromethane (DBCM) 266 216
Chlorodibromoacetic acid (CDBAA) 250 223
45
Discussion This research confirms that geosmin and 2-MIB were mainly reduced by the reaction with
hydroxyl radical (Rosenfeldt et al. 2005). Under UV/H2O2 conditions that provide substantial
removal of these odorants, brominated DBPs were also shown to be substantially removed.
However, chlorinated DBPs were not substantially removed under these conditions. For
halogenated DBPs such as THMs and HAAs, a possible first reaction step in UV/H2O2 could
be the hydrogen abstraction by hydroxyl radical or carbon-halogen bond cleavage by direct
UV photolysis. Bond dissociation energies (BDE) of the C-H bond in trichloromethane
(CHCl3) and tribromomethane (CHBr3) are very close; 100.0 and 99.9 kcal/mol, respectively
(McGivern, Derecskei-Kovacs et al. 2000). Thus BDE of C-H bond cannot explain the faster
removal of brominated DBPs. Carbon-bromine cleavage due to UV photolysis is the likely
mechanism of faster removal of brominated DBPs. This is supported by the higher strengths
of C-Cl bond than C-Br bonds in trichloromethane and tribromomethane, 80.1 and 70.4
kcal/mol respectively (McGivern, Derecskei-Kovacs et al. 2000), and higher molar
absorption coefficients of brominated DBPs (Figure 3-2).
Brominated DBPs were removed faster in proportion to the number of bromine atoms in
their structure. Tribromomethane and tribromoacetic acid (TBAA) were reduced the fastest
among the THMs and HAAs respectively. This can be explained by the fact that the C-Br
bond is the chromophore in these molecules; note that the molar extinction coefficient for
these compounds increases with increasing number of bromines. Regarding UV photolysis of
tribromomethane, water-catalyzed mechanism was proposed, in which isotribromomethane
recombinated from tribromomethane reacts with water molecule resulting O-H insertion and
HBr elimination (Li, Kwok et al. 2004). Consequently, in this research, brominated THMs
and HAAs were shown to be reduced by mainly UV photolysis in UV/H2O2 process.
Conclusion Many water treatment plants (WTPs) experience earthy-musty odor episodes during the
warm summer and fall months due to proliferation of cyanobacteria and production of
geosmin and 2-MIB. DBPs are typically at their highest level in the warm weather of summer
as well. Brominated DBPs are known to be more toxic than chlorinated DBP analogues, and
constitute about 10-40% of total DBPs produced. Therefore, UV/H2O2 process, when
46
implemented for odor control, can have the additional benefit of DBP reduction, especially in
the region where source water bromide concentration leads to brominated DBPs. This
simultaneous reduction of DBPs when applying UV/H2O2 to control earthy-musty odorants
has a substantial implication that UV/H2O2 can be seasonally used for controlling both
earthy-musty odorants and brominated DBPs.
Acknowledgement The authors specially thank Kwater (Korea Water Resources Corporation) for research
fellowship and MILES (Macromolecular Interfaces with Life Science) program in Virginia
Tech (National Science Foundation agreement # : DGE-0333378) for the experimental
support. The views expressed in this report are those of the authors and not those of the US
National Science Foundation.
47
Chapter 4. Reaction Mechanism of Haloacetic acid Degradation
in UV/H2O2 Advanced Oxidation Process
Abstract Haloacetic acids (HAAs) are class of the regulated disinfection byproducts (DBPs) in
drinking water. In previous research, brominated HAAs were shown to be reduced faster than
chlorinated DBPs by direct UV photolysis while chlorinated DBPs were reduced by hydroxyl
radical reaction in UV/H2O2. In this research, the removal mechanisms of HAAs in UV/H2O2
process were investigated using low pressure UV lamps of 253.7nm wavelength and 7.2
mW/cm2 total intensity and 100mL quartz reactor.
More bromine substituted HAAs had higher apparent pseudo-first order rate constant of
UV photolysis while less chlorine substituted HAAs had higher second order rate constants
of hydroxyl radical reaction. The moles of H+ and Br- or Cl- produced from the UV photolysis
of brominated HAAs and hydroxyl radical reaction of chlorinated HAAs were proportional to
the number of halogen atoms in HAAs. Two carbons in a HAA molecule were completely
mineralized with molar decrease of chlorinated HAAs while partial mineralization was
observed for brominated HAAs. Based on the postulated reaction mechanism, molar increase
ratio of hydrogen ion to halogen ion produced from both brominated and chlorinated HAAs
were 0, 0.5, 0.7 for mono-, di-, tri-halogenated HAAs, respectively, which was similar to
measured ratio of chlorinated HAAs. For MBAA and DBAA, molar increase ratio of H+ to
Br- produced from the UV photolysis of brominated HAAs were close to 1 due to incomplete
mineralization.
The C-Br bond cleavage is thought to be the first step of brominated HAAs degradation by
UV photolysis, followed by the reaction with oxygen or with water molecule. Faster removal
rates of brominated HAAs were associated with increased number of bromine atoms.
Hydrogen abstraction and electron transfer reaction are two possible first steps of the
degradation of chlorinated HAAs by hydroxyl radical. The stability of the transition state in
hydrogen abstraction and different electron density in electron transfer reaction can explain
faster removal of less chlorine substituted HAAs. The different reaction rates and removal
mechanisms of brominated and chlorinated HAAs indicate that UV/H2O2 oxidation will not
uniformly remove all HAA compounds.
48
Introduction Advanced oxidation process (AOP) involving hydroxyl radical (·OH) are applied to
remove organic contaminants from water. UV/H2O2 process is an AOP that produces
hydroxyl radical via the photolysis of H2O2; this process can efficiently remove organic
contaminants from water (Beltran, Ovejero et al. 1993; Stefan, Hoy et al. 1996; Stefan and
Bolton 1998; Cater, Stefan et al. 2000; Stefan, Mack et al. 2000; Rosenfeldt, Melcher et al.
2005; Rudra, Thacker et al. 2005; Paradis and Hoffman 2006; Rosenfeldt and Linden 2007).
Recently, UV/H2O2 has been applied in several full scale water treatment plants (WTPs) to
control earthy-musty odors from geosmin (trans-1,10-dimethyl-trans-9-decalol) and 2-MIB
(2-methylisoborneol), the disinfection byproduct, N-nitrosodimethylamine (NDMA), and the
industrial chemical, 1,4-dioxane (Cotton and Collins 2006).
The reaction mechanisms of UV/H2O2 consists of both direct UV photolysis and hydroxyl
radical reaction. Hydroxyl radical produced from the UV photolysis of hydrogen peroxide
plays a key role in UV/H2O2 process by oxidizing contaminants via radical chain reactions. In
case of compounds that have substantial UV absorbances, direct UV photolysis can mainly or
partially contribute to the removal of organic compounds in the UV/H2O2 process. In selected
research, UV irradiation without adding hydrogen peroxide has been shown to partially
remove geosmin, 2-MIB, diazinon, and mycrocystin (Qiao, Li et al. 2005; Rosenfeldt,
Melcher et al. 2005; Shemer and Linden 2006).
Many utilities are challenged by production of regulated disinfection byproducts (DBPs),
and their control frequently conflicts with fulfilling disinfection capacity required to
inactivated microorganisms such as Giardia and Cryptosporidium. Haloacetic acids (HAA)
are one of the typically regulated DBPs in drinking water; compared to THMs, they have
not been extensively studied (Buchanan, Roddick et al. 2006). Selected HAAs are known to
be more harmful to humans than THMs (Singer 2002). Although chlorinated HAAs are the
most prevalent, brominated HAAs are present in many source waters around the world
(Richardson, Thruston et al. 2003), and brominated HAAs were reported to typically consist
of 9-13% of total HAAs in U.S (Krasner, McGuire et al. 1989) and can locally consist of up
to 25% of total HAAs (Hyun, Kim et al. 2005; Buchanan, Roddick et al. 2006). Total HAA in
waters containing 0.1 mg/L bromide were reported to comprise at least 10% of brominated
HAAs (Cowman and Singer 1996). Brominated DBPs are known to be more toxic than their
chlorinated analogues (Echigo, Itoh et al. 2004; Richardson, Plewa et al. 2007) and can be
49
locally problematic in regions with higher aqueous bromide concentrations such as the
coastal areas or coal mining regions (von Gunten and Hoigne 1994; Richardson, Thruston et
al. 2003).
In regard to DBP, the AOP is known to reduce disinfection byproduct formation potential
(DBPFP) by breaking or changing the structures of precursors (Kusakabe, Aso et al. 1990;
Wang, Hsieh et al. 2000; Murray and Parsons 2004). The effect of UV/H2O2 on the formation
of THMs and HAAs were water specific in a bench-scale study (Paradis and Hoffman 2006).
However, to obtain higher UV transmittance, the AOP is generally used after filtration in full
scale treatment system. In addition, pre-chlorination is commonly used before coagulation in
WTPs, which means DBPs are already formed before AOPs are implemented.
However, there have not been sufficient studies on the removal of DBPs by AOPs and the
mechanisms. THMs were reported to be reduced at high UV intensity and H2O2 dose (17,000
mJ/cm2 and 0.1% respectively) (Rudra et al. 2005). It was reported that more bromine
substituted THMs have slightly greater reaction rate constants with hydroxyl radical than
chlorine substituted THMs (Mezyk, Helgeson et al. 2006). More bromine substituted THMs
were also shown to be better photolysed by UV (Nicole, De Laat et al. 1991). In a recent
study with UV/H2O2, brominated and chlorinated DBPs were shown to have different
removal mechanisms. Brominated DBPs were shown to be reduced faster than chlorinated
DBPs due to direct UV photolysis while chlorinated DBPs were reduced by hydroxyl radical
reaction. Consequently, UV/H2O2 was shown to significantly remove brominated DBPs at the
dose effective for removing geosmin and 2-MIB (Jo, Dietrich et al. 2008). Based on those
results, UV/H2O2 was thought to be one desirable oxidation process for the water of high
bromine concentration because ozonation causes bromate formation in high bromide source
waters.
In regard to reaction mechanisms of DBP degradation by UV photolysis or advanced
oxidation process, only a few mechanisms have been proposed. The UV photolysis
mechanism of tribromomethane (CHBr3) in aqueous solution was proposed to be a water-
catalyzed dehalogenation where HBr and CO2 were the final products (Li, Kwok et al. 2004).
Reaction mechanisms for other THMs have not been reported. In the case of haloacetic acids,
two pathways of dichloroacetic acid degradation with hydroxyl radical were proposed, where
two moles of chloride, hydrogen ion, and carbon dioxide were produced from a molar
decrease of dichloroacetate via hydrogen abstraction and electron transfer reaction (Zalazar,
Labas et al. 2007). Reaction mechanisms for haloacetic acids other than dichloroacetic acid
50
have not been reported. Reaction rate constants of halogen substituted acetic acids with UV
photolysis or hydroxyl radical reaction could be an important clue in investigating the
degradation mechanism of haloacetic acid in UV/H2O2 process and improving water
treatment and drinking water quality. UV photolysis reaction rates of brominated THMs and
HAAs were shown to increase with bromine substitution (Nicole, De Laat et al. 1991; Jo,
Dietrich et al. 2008). Reaction rate constants for chlorinated HAAs with hydroxyl radical
need to be further investigated. Reaction rate constants of chlorine substituted acetic acids
with several radical species exhibited an increasing trend of rate constants with fewer
chlorine substitutions. However, dichloroacetic acid had a greater rate constant with hydroxyl
radical than chloroacetic acid (Maruthamuthu, Padmaja et al. 1995). Therefore, it is needed to
further study on the reaction rate constant and mechanism of HAAs.
In this research, UV photolysis mechanism of brominated DBPs and hydroxyl radical
reaction mechanism of chlorinated DBPs in UV/H2O2 process were further studied. The
understanding of DBPs removal mechanism and kinetics of UV/H2O2 will be a useful
addition to the water treatment system design. The objectives of this research were 1) to
measure apparent pseudo-first order rate constants of UV photolysis of brominated HAAs
and second order reaction rate constants of chlorinated HAAs with hydroxyl radical, and 2) to
investigate the reaction mechanism of UV photolysis of brominated HAAs and hydroxyl
radical reaction of chlorinated HAAs in UV/H2O2 advanced oxidation process.
Methods and Materials
Apparatus UV lamps of 253.7 nm wavelength with 7.2 mW/cm2 total intensity (Rayonet RPR-100)
were used for UV irradiation system in the center of which a 100mL quartz vial was set as a
reactor. UV dose calculated by UV lamp intensity was verified with UV dose determined
from the iodide/iodate actinometer (Rahn 2004; Rahn, Bolton et al. 2006). Samples were
completely mixed by stirring while being irradiated with UV. A temperature of 22.5 ± 1.0℃
was maintained by an electric fan set blowing through dry ice and the UV system.
51
Reagents and Sample Preparation Six pure individual HAAs were diluted to desired concentrations in de-ionized water.
Higher concentrations than typical concentrations found in drinking water, by three order of
magnitude, were used to measure reaction rate constants and investigate reaction mechanisms
(Table 4-1). The HAAs tested consist of three chlorinated and three brominated compounds.
Chlorinated HAAs were chloroacetic acid (MCAA) (≥99%, Aldrich), dichloroacetic acid
(DCAA) (≥99%, Sigma-Aldrich), and trichloroacetic acid (TCAA) (≥99%, Alfa Aesar).
Brominated HAAs were bromoacetic acid (MBAA) (≥99%, Sigma-Aldrich), dibromoacetic
acid (DBAA) (≥99%, Fluka), and tribromoacetic acid (TBAA) (≥99%, Acros Organics).
Hydrogen peroxide (30%, Fisher) was diluted to desired concentrations of 6 - 12 mg/L, and
added into the samples immediately before UV irradiation. Tribromomethane (≥99%, Acros
Organics) and trichloromethane (≥99.9%, Fisher Scientific) are used as reference compounds
at concentrations of 82 µg/L and 120 µg/L for brominated HAAs and chlorinated HAAs,
respectively. Deuterated MCAA (MCAA-d3, 98% deuterium, Signma-Aldrich) was used to
investigate the isotope effect.
Table 4-1. concentrations of HAA compounds examined
Concentration
reaction rate constant experiment
mass balance experiment
HAAs
µg/L µM mg/L mM
pKa*
Chloroacetic acid (MCAA)
Dichloroacetic acid (DCAA)
Trichloroacetic acid (TCAA)
Bromoacetic acid (MBAA)
Dibromoacetic acid (DBAA)
Tribromoacetic acid (TBAA)
108
105
98
202
190
161
1.14
0.81
0.60
1.45
0.87
0.54
20
20
20
10
10
10
0.21
0.16
0.12
0.07
0.05
0.03
2.80
1.48
0.70
1.10
0.60
-
* (Nikolaou, Golfinopoulos et al. 2002)
Analysis HAAs were determined by liquid-liquid extraction method (EPA method 552.2) with a HP
5890 gas chromatograph (Avondale, PA, USA) and ECD detector (U.S. EPA 1995). Injector
temperature was 210℃ and initial oven temperature was set to 35℃ and ramped to 140℃.
UV absorbances were measured at a wavelength of 253.7 nm to calculate molar extinction
52
coefficients on a UV/Vis spectrophotometer (Beckman DU640). Bromide ion (Br-) and
chloride ion (Cl-) concentrations were measured by ion chromatography (Dionex DX-120)
based on EPA method 300.0 B. Brominated HAA concentration and bromide ion
concentration were also analyzed by UV absorbance where the parent brominated HAAs and
bromide ion concentrations were measured at a wavelength of 254 nm and 194 nm,
respectively (Li, Kwok et al. 2004; Zhao, Lin et al. 2005). In order to quantify mineralization,
TOC reduction was measured using a TOC analyzer (Sievers 800). Water pH was measured
by pH meter (Fisher Accumet 910) to quantify the increased concentration of hydrogen ion
released from the reaction. H2O2 concentration was determined by iodide (I3-) method
(Klassen, Marchington et al. 1994; Rosenfeldt, Melcher et al. 2005). Tribromomethane and
trichloromethane were measured based on Standard Method 6232.D by purge/trap (Tekmar
3000) and GC (Tremetrics 9001) with DB-624 column (J & W) (AWWA, APHA et al. 2005).
Experimental Procedure An appropriate concentration of aqueous HAA, with or without H2O2 was added to a 100
mL quartz reactor that was filled with the solution headspace free. The solution was de-
ionized water with an equilibrium amount of O2 from contact with room air. The initial pH of
de-ionized water was in the range of pH 6.60 to 6.98. The reactor containing the HAA was
irradiated for a predetermined time period that ranged from 10 seconds and 30 minutes. After
irradiation, samples were removed for analysis of HAA, bromide and chloride ion, TOC, and
pH. All samples were analyzed within 24 hours.
Results From the previous research, brominated HAAs were shown to be removed by UV photolysis,
not by hydroxyl radical reaction in UV/H2O2. On the other hand, chlorinated HAAs were
shown to be removed by hydroxyl radical reaction (Jo, Dietrich et al. 2008). In this research,
removal mechanism of HAA was investigated by measuring reaction rate constants and
making balances of hydrogen ion, halogen ion, and carbon.
Apparent reaction rate constant of brominated HAAs Apparent pseudo-first order reaction rate constants of UV photolysis of brominated HAAs
were derived by measuring slopes in the plot of log removal (ln[C/C0]) versus time (Figure 4-
53
1). Reaction rate constants of UV photolysis of three brominated HAAs were compared to
each other and tribromomethane, which was shown to decrease the fastest among the
brominated and chlorinated THMs in previous studies (Nicole, De Laat et al. 1991). Like
THMs, more bromine substituted HAAs were removed faster in UV photolysis.
Tribromoacetic acid was shown to have twice the reaction rate constant of tribromomethane.
Dibromoacetic acid had a slightly lower reaction rate constant than tribromomethane (Table
4-2). This faster removal of more bromine substituted HAAs results from more C-Br
chromophores in their molecule in previous research (Jo et al. 2008). The relative ratios of
measured apparent reaction rate constants of brominated HAAs were 1:8:34, which was
greater than the expected ratio (1:2:3). This greater difference in reaction rates of brominated
HAAs can be explained by the ratio of molar extinction coefficients which was 1:5:24 as
shown in Figure 4-2.
Figure 4-1. Removal rate of brominated HAAs and tribromomethane by UV photolysis
Figure 4-2. Molar absorption coefficients measured at 254 nm in this research
54
Table 4-2. Apparent pseudo-first order reaction rate constants for UV photolysis of three
brominated HAAs and tribromomethane
Compound Apparent k measured (s-1) Relative rate
constant versus MBAA
Bromoacetic acid (MBAA) 1.9 x 10-3 1
Dibromoacetic acid (DBAA) 1.5 x 10-2 8
Tribromoacetic acid (TBAA) 6.3 x 10-2 34
Tribromomethane 2.4 x 10-2 13
Reaction mechanism of brominated HAAs in UV/H2O2 process To investigate the mechanisms, molar increases of bromide and hydrogen ion concentration
were plotted with molar decrease of HAA. Molar TOC removal was plotted with molar HAA
removal to determine whether complete mineralization takes place or stable intermediates are
present. This approach was based on previous studies on the reaction rate and mechanism of
trihalomethane and DCAA (Nicole, De Laat et al. 1991; Li, Kwok et al. 2004; Zalazar, Labas
et al. 2007).
[H+]and [Br-] balance
[H+] and [Br-] produced from the UV photolysis of three brominated HAAs were measured
and plotted (Figure 4-3). The ratio of molar increase in [Br-] from MBAA to molar increase
in [Br-] from DBAA and TBAA was 1:1.7:2.9 which was close to the theoretical ratio, 1:2:3.
Ratios of [H+] to [Br-] were almost 1:1 for each of the three compounds even though less [H+]
was consistently produced than [Br-] for three brominated HAAs. The experiment was
repeated three times and results were similar (data not shown).
This result indicates that the moles of H+ and Br- produced from the UV photolysis of molar
brominated HAAs was in proportion to the number of bromine atoms in HAAs. Bromide ion
production is thought to result from the C-Br bond cleavage because the C-Br bond is the
chromophore, and molar extinction coefficient increases with the number of C-Br bonds.
Hydrogen ion is thought to be released from photo-assisted hydrolysis of the molecule.
55
Figure 4-3. Molar increase of [H+] and [Br-] with molar decrease of brominated HAAs
exposed to UV photolysis at 253.7 nm wavelength (Regression equations are provided for
bromide ion)
Carbon balance
To investigate mineralization, a carbon balance was made by measuring total organic carbon
(TOC). Two moles of TOC were completely removed with a molar decrease of TBAA.
However, only 10% and 30% of TOC were removed for MBAA and DBAA, respectively
(Figure 4-4). This result indicates that the two carbons in a TBAA molecule were completely
mineralized while two carbons in MBAA and DBAA molecules were partially mineralized
under the condition used in this research and subsequently, stable intermediate(s) would be
present in the UV photolysis of MBAA and DBAA .
56
Figure 4-4. Molar decrease of TOC with molar decrease of HAA concentration for three
brominated HAAs
Postulated UV photolysis mechanism of brominated HAAs
Based on the results above, UV photolysis mechanisms of brominated HAAs were
examined. Result of [H+] and [Br-] balance indicates that H+ and Br- would be produced
according to the number of bromine atoms in HAA molecules. Carbon balances indicate that
there were partial mineralizations for MBAA and DBAA. In the case of mineralization, two
moles of carbon dioxide would be produced as a final product. In contrast, incomplete
mineralization would produce stable intermediates resulting in different pathways. C-Br bond
cleavage was thought to be the first step in this UV photolysis because C-Br bond is the
chromophore in the HAA molecule.
In the case of mineralization, two mechanisms are possible. From the first step, two radicals
can be produced (·CXnHn-3COO- and ·Br). The second step can be a reaction with oxygen.
With a presence of oxygen, the ·CXnHn-2COO- radical is likely to react with oxygen
producing peroxyl radical, ·OOCXnHn-2COO- as frequently seen in other radical reactions
(Fliount, Makogon et al. 1997; Makogon, Fliount et al. 1998; Zalazar, Labas et al. 2007)
(pathway 1). Another possible explanation is O-H insertion/H-Br elimination resulting from
the interaction with water molecule (pathway 2). Based on these results, when complete
mineralization takes place, the reaction mechanisms of brominated HAAs were postulated as
follows:
57
MBAA-pathway 1
CH2BrCOO- + hν → ·CH2COO- + ·Br
·Br + e- → Br-
·CH2COO- + O 2 → ·OOCH2COO-
·OOCH2COO- → -OOĊH2 + CO2 -OOĊH2 → CO2 +2H+ + 3e-
O 2 + 2H+ + 2e- → H2O2
----------------------------------------------------------
Overall reaction CH2BrCOO- + hν + 2O 2→ 2CO2 + Br- + H2O2
MBAA-pathway 2
CH2BrCOO- + hν + H2O → HOCH2COO- + H+ + Br-
HOCH2COO- + H2O → 2CO2 +5H+ + 6e-
2O 2+ 6H+ + 6e-→ 2H2O + H2O2
----------------------------------------------------------------
Overall reaction CH2BrCOO- + hν + 2O 2→ 2CO2 + Br- + H2O2
DBAA-pathway 1
CHBr2COO- + hν → ·CHBrCOO- + ·Br
·Br + e- → Br-
·CHBrCOO- + O 2→ ·OOCHBrCOO-
·OOCHBrCOO- → -OOĊHBr +CO2 -OOĊHBr → CO2 + Br- + H+ + e-
----------------------------------------------------------------
Overall reaction CHBr2COO- + hν + O 2→ 2CO2 + 2Br- + H+
DBAA-pathway 2
CHBr2COO- + hν + H2O → HOCHBrCOO- + H+ + Br-
HOCHBrCOO- + H2O → 2CO2 + Br- + 4H+ + 4e-
O 2+ 4H+ + 4e-→ 2H2O
-------------------------------------------------------------------
Overall reaction CHBr2COO- + hν + O 2→ 2CO2 + 2Br- + H+
58
TBAA-pathway 1
CBr3COO- + hν → ·CBr2COO- + ·Br
·Br + e- → Br-
·CBr2COO- + O 2→ ·OOCBr2COO-
·OOCBr2COO- → -OOĊBr2 +CO2 -OOĊBr2 + 2H2O → CO2 + 2Br- + 2H+ + H2O2 + e-
----------------------------------------------------------------------
Overall reaction CBr3COO- + hν + 2H2O+ O 2→ 2CO2 + 3Br- + 2H+ + H2O2
TBAA-pathway 2
CBr3COO- + hν + H2O → HOCBr2COO- + H+ + Br-
HOCBr2COO- + H2O → 2CO2 + 2Br- + 3H+ + 2e-
O 2 + 2H+ + 2e-→ H2O2
-----------------------------------------------------------------------
Overall reaction CBr3COO- + hν + 2H2O+ O 2→ 2CO2 + 3Br- + 2H+ + H2O2
Based on these postulated reaction mechanisms, less hydrogen ion than bromide ion would
be produced, and no hydrogen ion increase would be observed from the reaction of MBAA.
However, almost same amounts of H+ compared to Br- were produced from the three
brominated HAAs (Figure 4-3). This can be explained by the incomplete mineralization
which were observed in the carbon balance of MBAA and DBAA (Figure 4-4). In the case of
incomplete mineralization, organic intermediate and hydrogen ion can be produced. One
possible explanation is the formation of oxalate. Oxalate formation causes incomplete
mineralization because oxalate is detected as an organic compound in the TOC measurement.
With the oxalate production, hydrogen ion would be produced from the following reaction:
CH2BrCOO- + hν + O 2 → HC2O4- + Br- + H+
CHBr2COO- + hν + O 2 + 2H2O → HC2O4- + 2Br- + 2H+ + H2O2
To investigate this assumption, percentage of mineralization and oxalate formation was
calculated respectively from the measured ∆TOC/∆HAA, and measured ∆[H+]/∆HAA was
compared to calculated ∆[H+]/∆HAA that was determined by mineralization percentage and
theoretical ratio of ∆[H+]/∆HAA (Table 4-5). Measured ∆[H+]/∆HAA for the brominated
59
HAAs were similar to calculated ∆[H+]/∆HAA for MBAA and DBAA. Therefore, production
of oxalate was thought to be a possible explanation of the increased H+ production observed
in the UV photolysis of these two brominated HAAs compared to the amount of H+ expected
from the mineralization mechanism.
Table 4-3. Comparison of measured and calculated ∆[H+]/∆HAA based on percent
mineralization
Parameter MBAA DBAA TBAA
∆TOC/∆HAA Measured 0.2 0.6 2.1
Expected* 2.0 2.0 2.0
∆[H+]/∆HAA Measured 0.9 1.7 2.6
Expected* 0.0 1.0 2.0
∆[Br-]/∆HAA Measured 1.0 1.8 2.9
Expected* 1.0 2.0 3.0
% incomplete mineralization** (HC2O4
- formation) 88.5 79.1 0.0
∆[H+]/∆HAA for incomplete mineralizaiton 1.0 2.0 3.0
% mineralization** (CO2 formation) 11.5 20.9 100.0
∆[H+]/∆[HAA] for complete mineralization 0.0 1.0 2.0
Calculated ∆[H+]/∆HAA *** 0.9 1.8 2.0
*based on complete mineralization **calculated from measured ∆TOC/∆HAA (when ∆TOC/∆HAA = 2, complete mineralization = 100%) ***calculated from the sum of (percentage times theoretical ∆[H+]/∆[HAA]) for incomplete and complete mineralization
Reaction rate constants of Chlorinated HAAs and Hydroxyl Radicals Second order reaction rate constants of chlorinated HAAs and hydroxyl radicals were
measured by competition kinetics using trichloromethane as a reference compound (Figure 4-
5). UV at 253.7 nm wavelength and 6 mg/L H2O2 were used as a hydroxyl radical source.
Second order reaction rate constants of chlorinated HAAs with hydroxyl radicals can be used
to explain the chlorine substitution effect on the hydroxyl radical reaction rate. In addition,
60
those reaction rate constants provide some clue about the reaction mechanism such as the role
of hydrogen abstraction. Reaction rate constants measured in this research were similar to
those for chlorinated HAAs reported by other researchers (Adams, Boag et al. 1965;
Yokohata, Ohmura et al. 1969; Maruthamuthu, Padmaja et al. 1995). However, based on the
results, less chlorine substituted HAAs had higher reaction rate constant (Figure 4-5 and
Table 4-4). This is the opposite to the faster removal of more bromine substituted HAAs that
was shown to be degraded by C-Br bond cleavage resulting from UV photolysis. Considering
two possible pathways, hydrogen abstraction and electron transfer reaction, as shown in the
equation (1) and (2), higher reaction rate constants of less chlorine substituted HAAs can be
explained by two mechanisms.
CClH2COO- + ·OH → ·CClHCOO- + H2O Hydrogen abstraction (1)
CH2ClCOO- + ·OH → CH2ClCOO· + HO- Electron transfer (2)
Figure 4-5. Removal rates of three chlorinated HAAs compared to trichloromethane
61
Table 4-4. Second order reaction rate constants of chlorinated HAAs
Compound Measured k in this research (M-1s-1)
Relative rate constant versus
TCAA Reported k in literature (M-1s-1)
MCAA 3.3 x 108 46 8.3 x 107 a
4.0 x 108 b
4.3 x 107 c
DCAA 1.5 x 108 21 1.0 x 108 a
TCAA 7.2x 106 1 6.0 x 107 a * a (Maruthamuthu, Padmaja et al. 1995) b (Yokohata, Ohmura et al. 1969) c (Adams, Boag et al. 1965) * This value was reported as an upper limit due to impurity issue in the research
Faster removal of MCAA and DCAA than TCAA implies that hydrogen abstraction is a
likely the first step in this reaction, because there are two, one, and zero abstractable
hydrogen atoms in their molecules, respectively. The rate of TCAA removal is also less than
trichloromethane, which has an abstractable hydrogen atom. Deuterated MCAA was
compared to MCAA for the reaction rate with hydroxyl radical to assess the isotope effect.
Reaction of deuterated MCAA with hydroxyl radical was slower than MCAA, and the isotope
effect (kH/kD) was 2.9 indicating that hydrogen abstraction takes place as a rate-limiting step
(Figure 4-6).
Figure 4-6. Comparison of reaction rates with UV/H2O2 between deuterated MCAA and MCAA
62
Faster removal of MCAA than DCAA can be explained by the stability of the transition state
in the hydrogen abstraction step. In the transition state, the hydroxyl radical has a partial
negative charge having an interaction with an abstractable hydrogen atom. On the other hand,
the abstractable hydrogen atom has a partial positive charge which makes its transition state
more unstable when the carbon atom also has a less electron density induced by
electronegativity of substituted chlorine causing partial positive charge (Figure 4-6 and 4-7).
Therefore, hydrogen abstraction of MCAA that has only one chlorine atom is more favorable
than DCAA because halogen substituted carbon has more electron density in the transition
state of MCAA due to less electron withdrawing of a single chlorine atom.
Figure 4-7. Transition state for hydrogen abstraction of DCAA; both chlorine atoms withdraw
electron density from the carbon atom and destabilize the transition state
Figure 4-8. Transition state for hydrogen abstraction of MCAA
In the case of an electron transfer reaction, another explanation for the faster removal of
MCAA is by the stability of the acetate ion and electron density around the carboxylic carbon.
Compared to TCAA, MCAA has fewer chlorine atoms in the molecule, which makes the
chloroacetate ion less stable due to less partial positive charge on the carbon atom with the
63
lower electronegativity of chlorine (Figure 4-9). Subsequently, in an electron transfer reaction,
chloroacetate has higher electron density around the carboxylic oxygen, which makes
electron transfer from carboxyl group to hydroxyl radical easier causing faster removal. This
is consistent with the fact that MCAA has the highest pKa among the three chlorinated HAAs
(Table 4-1).
Figure 4-9. Partial positive charge on the chlorinated carbon atom of acetate ion; the partial
positive charge on the carbon of chloroacetate is less than in trichloroacetate ion
Reaction mechanism of chlorinated HAAs in UV/H2O2 process
[H+] and [Cl-] balance
Chloride ion was produced in proportion to the number of chlorine atoms in a chlorinated
HAA molecule. The ratio for molar increase of chloride ion to decrease of chlorinated HAA
was 1:2.5:3.1 for MCAA, DCAA, and TCAA, respectively (Figure 4-10). Less hydrogen ion
was released from the reaction than chloride ion. From the reaction of MCAA, hydrogen ion
was barely produced. For DCAA and TCAA, ratios of increased molar hydrogen ion
concentration to chloride ion were 0.4 and 0.5, respectively.
64
Figure 4-10. Molar increase of [H+] and [Cl-] compared to molar decrease of corresponding
chlorinated HAA
Carbon balance
Molar decrease ratio of TOC to three chlorinated HAAs were close to expected ratio of 2:1
based on total mineralization (Figure 4-11). According to this carbon balance, two carbons in
a HAA molecule were completely mineralized and no stable intermediates were present in the
reaction of chlorinated HAAs with hydroxyl radical.
Figure 4-11. Molar decrease of TOC with molar decrease of three chlorinated HAAs
65
Two pathways are possible for the first step of the reaction mechanism; hydrogen
abstraction and electron transfer reaction. According to the faster removal rate of less chlorine
substituted HAAs and observed isotope effect, hydrogen abstraction is indicated to be a
reaction mechanism except for TCAA which has no abstractable hydrogen atom. Electron
transfer which was the only pathway for TCAA was also able to explain the faster removal of
less chlorine substituted HAAs and lead to same overall reaction. These results were
consistent with the previous research where both hydrogen abstraction and electron transfer
were proposed as the first step of the hydroxyl radical reaction of DBAA (Zalazar, Labas et al.
2007). Base on the results, hydroxyl radical reaction mechanisms of chlorinated HAAs were
postulated as follows:
MCAA mechanism 1 - Hydrogen abstraction
1/2H2O2 → ·OH
CClH2COO- + ·OH → ·CClHCOO- + H2O
·CClHCOO- + O2 → ·OOCClHCOO-
·OOCClHCOO- → 2CO2 + H+ + Cl- + e-
H+ + e- + 1/2O2 → 1/2H2O2
----------------------------------------------------------
Overall reaction CClH2COO- + 3/2O2 → 2CO2 + Cl- + H2O
MCAA mechanism 2 � Electron transfer
1/2H2O2 → ·OH
CH2ClCOO- + ·OH → CH2ClCOO· + HO-
CH2ClCOO· → ·CH2Cl + CO2
·CH2Cl + O2 → H2ClCOO·
H2ClCOO· + → CO2 + 2H+ + Cl- + e-
H+ + e- + 1/2O2 → 1/2H2O2
----------------------------------------------------------
Overall reaction CH2ClCOO- +O2 →2CO2 + Cl- + H2O
66
DCAA mechanism 1 � Hydrogen abstraction
1/2H2O2 → ·OH
CCl2HCOO- + ·OH → ·CCl2COO- + H2O
·CCl2COO- + O2 → ·OOCCl2COO-
·OOCCl2COO- + 2H2O → 2CO2 + 2Cl- + 2H+ + H2O2 + e-
H+ + e- + 1/2O2 → 1/2H2O2
-----------------------------------------------------------------------
Overall reaction CCl2HCOO- + 3/2O2+ H2O → 2CO2 + 2Cl- + H+ + H2O2
DCAA mechanism 2 � Electron transfer
1/2H2O2 → ·OH
CHCl2COO- + ·OH → CHCl2COO· + HO-
CHCl2COO· → ·CHCl2 + CO2
·CHCl2 + O2 → Cl2HCOO·
Cl2HCOO· + 2H2O → CO2 + 2Cl- + 3H+ + H2O2 + e-
H+ + e- + 1/2O2 → 1/2H2O2
-----------------------------------------------------------------------
Overall reaction CCl2HCOO- + 3/2O2+ H2O → 2CO2 + 2Cl- + H+ + H2O2
TCAA mechanism � Electron transfer
1/2H2O2 → ·OH
CCl3COO- + ·OH → CCl3COO· + HO-
CCl3COO· → ·CCl3 + CO2
·CCl3 + 3/2H2O2→ CO2 + 3Cl- + 3H+ + 1/2O2
----------------------------------------------------------------------
Overall reaction CCl3COO- + 2H2O2 →2CO2 + 3Cl- + 2H+ + H2O + 1/2O2
Based on postulated reaction mechanisms, no hydrogen ion would be produced from the
reaction of MBAA, and ratios hydrogen ion to chloride and are 0.5 and 0.7 for DCAA and
TCAA, respectively. This was consistent with the result where hydrogen ion barely increased
for MCAA, and higher ratio of hydrogen ion to chloride was observed from TCAA than
DCAA. In regard to the hydrogen ion balance, measured parameters were similar to expected
ones calculated from postulated reaction mechanisms (Figure 4-10). Consequently, postulated
67
mechanisms well explain the observed production of hydrogen and chloride ion.
Table 4-5. Comparison of measured and expected parameters of chlorinated HAAs
Parameter MCAA DCAA TCAA
Measured 2.1 2.1 2.2 ∆TOC/∆HAA
Expected* 2.0 2.0 2.0
Measured 0.1 1.0 1.5 ∆[H+]/∆HAA
Expected* 0.0 1.0 2.0
Measured 0.1 0.4 0.5 ∆[H+]/∆[Cl-]
Expected* 0.0 0.5 0.7 *Theoretical ratio based on postulated mechanism
Discussion
UV photolysis mechanism of brominated HAA
Although both brominated and chlorinated HAAs can be photodegraded in water (Lifongo,
Bowden et al. 2004), due to the difference in UV absorption, brominated compounds were
shown to be removed by either direct UV photolysis or hydroxyl radical reaction, while
chlorinated compounds were shown to only react with radical species (Fliount, Makogon et al.
1997; McGivern, Kim et al. 2004; Jo, Dietrich et al. 2008). According to the results, slightly
more bromide and less hydrogen ion were produced from the UV photolysis in proportion
with the number of bromine atoms in the brominated HAA molecules even though less
hydrogen ion production was expected based on postulated mineralization mechanism. Based
on observed incomplete mineralization of MBAA and DBAA, oxalate was proposed as a
possible reaction intermediate. Oxalate formation via incomplete mineralization was able to
explain more hydrogen ion production than the postulated complete mineralization
mechanism, which however, needs to be further studied. In aqueous phase, homolysis of C-Br
bonds of molar tribromomethane was reported to release three moles of bromide from the
compound and three moles of hydrogen ion from the water while the other part of water
molecule (O-H) react with the compound (Li, Kwok et al. 2004). This H-Br removal/O-H
insertion mechanism could be one possible mechanism of UV photolysis of brominated
HAAs. However, in many radical reactions of halogenated compounds, peroxyl radical
formation was reported to occur after C-X bond cleavage (Fliount, Makogon et al. 1997;
68
Makogon, Fliount et al. 1998; Li, Stefan et al. 2004; McGivern, Kim et al. 2004). Therefore,
the reaction with oxygen is more likely the second step. To elucidate the role of oxygen in
this removal, oxygen was removed from the solution by purging nitrogen or helium gas
before the UV irradiation. However, because realistic environmental µg/L concentrations of
HAAs were used in this research, the dissolved oxygen concentration could not be lowered
below the level that prevents peroxyl radical formation.
Hydroxyl radical reaction mechanism of chlorinated HAAs
Although hydrogen abstraction is thought to be one pathway according to the observed
isotope effect, electron transfer reaction is likely to be another important pathway in hydroxyl
radical reaction of chlorinated HAAs based on stability and electron density on carboxylic
oxygen of less chlorine substituted HAA. This is consistent with other studies where electron
transfer was proposed as one of two pathways of the hydroxyl radical reaction of DCAA
(Legrini, Oliveros et al. 1993; Zalazar, Labas et al. 2007). Faster removal of MCAA observed
in this research is also consistent with other study that reported Kolbe reaction following
electron transfer step was more effective for less halogenated HAAs due to higher electron
density at the carboxyl function (Mao, Schoeneich et al. 1991).
CCl2COO· → ·CHCl2 + CO2 (Kolbe mechanism)
Another consideration is the reaction of chloride and hydroxyl radical during the hydroxyl
radical reaction of chlorinated HAAs (Cl- + ·OH → Cl2-· + OH-). Maruthamuthu et al.
reported MCAA is removed faster likely because chloride reacts with hydroxyl radical
producing Cl2-·, which reacts faster with MCAA than DCAA (Maruthamuthu, Padmaja et al.
1995). Nevertheless, increased chloride ion concentrations measured in this research were
equal to or greater than expected, which indicates that the reaction of chloride ion and
hydroxyl radical has a minimal effect on pH change.
Future work
In some of the reaction mechanisms postulated in this research, hydrogen peroxide was
proposed as a final product. Concentration change of hydrogen peroxide needs to be
measured to confirm this mechanism in the future work. In the case of incomplete
mineralization of brominated HAAs, oxalate formation was proposed. To confirm oxalate
formation as an incomplete mineralization, oxalate concentration also needs to be measured
before and after UV photolysis in the future work.
69
Conclusion
Bromine substituted HAAs were photolysed producing hydrogen ion and bromide. TBAA
was completely mineralized under the condition used in this research while MBAA and
DBAA were partially mineralized. The C-Br bond cleavage is thought to be the first step
followed by the second step, the reaction with oxygen. Interaction with a water molecule is
possibly another second step. For MBAA and DBAA, more hydrogen ion was produced than
postulated mineralization mechanism, which was explained by the production of oxalate via
incomplete mineralization. More bromine substituted HAAs have greater reaction rates in
direct UV photolysis. These results can be explained by the more C-Br bond and higher
molar extinction coefficients.
Chlorine substituted HAAs were mineralized by hydroxyl radical reaction in UV/H2O2
process producing chloride and hydrogen ion. Molar increase ratios of chloride to decreased
chlorinated HAAs were proportional to the number of chlorine in the molecules. However,
unlike brominated HAAs, less chlorine substituted HAAs had greater second order reaction
rate constants. Both hydrogen abstraction and electron transfer reaction were thought to be
two first steps, and were able to explain the removal rate, hydrogen and chloride ion balance,
and carbon balance of chlorinated HAA.
Acknowledgement
The authors specially thank Kwater (Korea Water Resources Corporation) for the research
fellowship to support Jo and MILES (Macromolecular Interfaces with Life Science) program
in Virginia Tech (National Science Foundation agreement # : DGE-0333378) for the
experimental support. The views expressed in this report are those of the authors and not
those of the US National Science Foundation.
70
Chapter 5. Removal and Transformation of Odorous
Aldehydes by UV/H2O2
Abstract Removal of odorous aldehydes by UV/H2O2 was compared to that of geosmin and 2-MIB
by the same process. Odor transformation was investigated by sensory test and byproducts
were monitored by a carbonyl derivatization method. Heptadienal, decadienal, and
nonadienal were removed faster than geosmin and 2-MIB. The primary mechanism was the
direct UV photolysis in the UV/H2O2 process. In sensory tests, new odors such as chalky or
sweet odors were produced while the initial odor intensity of fishy/grassy-smelling aldehydes
was reduced with increasing exposure time to UV/H2O2. New carbonyl compounds were
detected from the UV photolysis of nonadienal and were not removed by further UV
irradiation, which was thought to be related with production of new odors. Results indicate
that new types of odor were produced from the oxidation of odorous aldehydes, and
consequently, sensory tests coupled with chemical analysis should be considered in designing
oxidation process to control recalcitrant odorants.
Keywords : Odor, aldehyde, UV photolysis, UV/H2O2, AOP
Introduction Recently, more interest has been focused on drinking water aesthetic issues. This trend
indicates that consumers demand �more pleasant� or �more tasty� drinking water as well as
safe water (Devesa, Fabrellas et al. 2004; Khiari 2004; Burlingame and Mackey 2007; Liang,
Wang et al. 2007). Consumer comparison of tap water to bottled water may intensify this
trend. Various efforts have been made to remove recalcitrant odorants in drinking water to
prevent complaints and meet consumer standards about drinking water quality. Many
researchers reported that advanced oxidation processes (AOPs), which involve hydroxyl
radical, efficiently reduce earthy/musty odorants (geosmin and 2-MIB) in drinking water
(Rosenfeldt, Melcher et al. 2005; Paradis and Hoffman 2006; Westerhoff, Nalinakumari et al.
2006; Jo, Dietrich et al. 2008). Odorous aldehydes such as nonadienal and heptadienal are
mostly produced from algae and can cause off-flavor in drinking water, especially in the case
71
of insufficient chlorination (Burlingame, Muldowney et al. 1992; Andersson, Forsgren et al.
2005). Nonadienal had a greater reaction rate constant with hydroxyl radical than geosmin
and 2-MIB (Peter and Von Gunten 2007). However, it was reported that some algal
metabolites were transformed into new types of odor by oxidation (Dietrich, Hoen et al.
1995), and fruity smelling aldehyde were produced from the ozonation of drinking water
(Anselme, Suffet et al. 1988; AWWARF 1995; Bruchet and Duguet 2004). Low molecular
weight aldehydes, which are possible product of oxidation of unsaturated aldehyde, were
considered to be related with off-flavor events (Fabrellas, Matia et al. 2004).
In this research, removal of odorous aldehydes by UV/H2O2 were compared to geosmin
and 2-MIB, and odor transformation was investigated by sensory test and pentafluorobenzyl-
hydroxylamine hydrochloride (PFBHA) derivatization method to detect carbonyls. The
UV/H2O2 process performs by direct UV photolysis and hydroxyl radical reaction (Cotton
and Collins 2006). Hydroxyl radical produced from the UV photolysis of hydrogen peroxide
plays a key role for many reactions. However, for the compounds that greatly absorb UV,
direct UV photolysis may be the main mechanism in the removal of the compounds by
UV/H2O2 process (Nicole, De Laat et al. 1991; Qiao, Li et al. 2005; Jo, Dietrich et al. 2008).
The objectives of the research were: 1) to compare removal rates of fishy/grassy smelling
aldehydes to geosmin/2-MIB in UV/H2O2 process, 2) to elucidate the main mechanism of
odorous aldehyde removal in UV/H2O2 process, 3) to investigate how odor intensities and
descriptors change during the reaction with UV/H2O2, and 4) to detect intermediates and final
products.
Materials and Methods Four types of aldehydes were selected from the typical algae-related fishy/grassy odorants
as well as geosmin and 2-MIB. Compounds used in this research were: trans-2,cis-6-
nonadienal (Aldrich, 92%, CAS no. 552-48-2), hexanal (Aldrich, 98%, CAS no. 66-25-1),
trans-2,trans-4-decadienal (TCI, 98%, CAS no. 25152-84-5), trans-2,trans-4-heptadienal
(TCI, 90%, CAS no. 4313-03-5), geosmin (Sigma, 98%, CAS no. 16423-19-1), 2-MIB
(Supelco, 99.9%, CAS no. 2371-42-8). Initial concentrations were selected based on
threshold and detection limit (Watson, Satchwill et al. 2001; Satchwill, Watson et al. 2007).
Structures and odor properties of these compounds are shown in Table 5-1. Experiments were
performed with a 253.7 nm wavelength UV lamp of 7.2 mW/cm2 intensity (Rayonet RPR-
100) with quartz reactors. H2O2 concentration of 6 mg/L was used considering the optimal
72
range of H2O2 dosage in previous research (Cotton and Collins 2006; Paradis and Hoffman
2006). Samples were prepared in de-ionized water (Nanopure) and completely mixed and
headspace free while being irradiated with UV. Odorants were dosed at µg/L concentrations
and measured by solid-phase microextraction (SPME, Supelco) with scan mode of GC/MS
(Agilent 5973) (Watson, Brownlee et al. 1999; Watson, Brownlee et al. 2000).` UV
absorbances were measured at a wavelength of 253.7 nm by UV/Vis spectrophotometer
(Beckman DU640). H2O2 concentration was determined by iodide (I3-) method (Klassen,
Marchington et al. 1994; Rosenfeldt, Melcher et al. 2005). UV dose was verified with the
iodide/iodate actinometer (Rahn 2004; Rahn, Bolton et al. 2006). Flavor Profile Analysis
(FPA) was performed by four trained panelists according to the Standard Method 2170 to
assess the odor intensity and investigate the change of odor descriptor (AWWA, APHA et al.
2005). In FPA, panelists smelled 8 samples per session including odor free sample, and
discussed on the odor descriptors and intensities. One or two sessions were held for one
compound coupled with chemical analysis. PFBHA derivatization method was used with
SPME and GC/MS to detect low molecular weight carbonyl groups (aldehydes and ketones)
produced from the oxidation of nonadienal (Weinberg and Glaze 1997; Bao, Pantani et al.
1998), where higher concentration (10 mg/L) of nonadienal were reacted by UV/H2O2 and
subsequently derivatized with PFBHA.
Table 5-1. Odorants selected for this research
Compounds Structure Odor Odor threshold (ng/L)
Guideline in drinking
water trans-2,cis-6-
nonadienal
O
Cucumber/Fishy 80 a -
trans-2,trans-4-
decadienal
O
Fishy/Oily/Cucumber 300 b -
trans-2,trans-4-
heptadienal
O
Grassy/Oily/Fishy 25,000 b -
Hexanal O
Grassy/Sweet 4,500 c -
Geosmin
Earthy 6-10 d 10 ng/L e
2-MIB
Musty 2-20 d 10 ng/L e
a (Young, Horth et al. 1996)
73
b (Watson, Satchwill et al. 2001) c (Rychlik, Schieberle et al. 1998) d (Rashash, Dietrich et al. 1997; Oestman, Schweitzer et al. 2004) e Guideline in Korea and secondary standard in Japan (KNIER 2000; KMOE 2006)
Results
UV absorbance
In order to assess the contribution of direct UV photolysis, molar extinction coefficients
which indicate the UV absorbance of a compound were measured as shown in Figure 5-1.
Three unsaturated aldehyde compounds absorbed greater amount of UV compared to
geosmin and 2-MIB. The order of molar extinction coefficient from greatest to least was
heptadienal, decadienal, and nonadienal. Based on the measured molar extinction coefficients,
it was expected that three unsaturated aldehyde compounds would be reduced much faster
than geosmin and 2-MIB by UV photolysis. In contrast, UV absorbance of hexanal and
decanal were almost zero, which indicates that removal of these compounds, if any, would be
by hydroxyl radical reaction in the UV/H2O2 process.
Figure 5-1. Molar extinction coefficient measured in this research (M-1cm-1)
74
Removal rate by UV/H2O2
Compared to geosmin and 2-MIB, the three �dienal� compounds were removed faster.
Heptadienal was reduced faster than either nonadienal or decadienal, which is thought to be
related to its higher UV absorbance. Nonadienal and decadienal had similar removal rates to
each other. Hexanal was not better removed than geosmin.
Figure 5-2. Log removal of odorants with UV dose (6 mg/L H2O2 )
Sensory test
Sensory tests revealed that the initial odor intensity of odorous aldehydes was reduced with
increasing exposure time to UV/H2O2. However, new types of odors were detected when the
initial fishy/grassy odors were mostly or completely removed. Fishy/cucumber odor of
nonadienal changed into sweet/chalky odor (Figure 5-3) as concentration of nonadienal was
reduced by UV/H2O2. This sweet/chalky odor was thought to be produced from the oxidation
of nonadienal. Oily/fishy/cucumber odor of decadienal changed into sweet/stale odor (Figure
4). Grassy/oily/fishy odor of heptadienal changed into sweet/concrete/wet cardboard odor
(Figure 5-5). Grassy/sweet/pumpkin odor of hexanal changed into cement/waxy/metallic/oily
odor (Figure 5-6). Consequently, in the oxidation of odorous �dienal� compounds by
UV/H2O2, new types of odors were produced as the concentration of the original compounds
and initial odors were reduced. These results indicate that the oxidation of odorous aldehyde
by UV/H2O2 produce byproducts that have different types of odor.
75
Figure 5-3. Nonadienal concentration and odors as a function of UV dose (6 mg/L H2O2)
Figure 5-4. Decadienal concentration and odors as a function of UV dose (6 mg/L H2O2)
Figure 5-5. Heptadienal concentration and odors as a function of UV dose (6 mg/L H2O2)
76
Figure 5-6. Hexanal concentration and odors as a function of UV dose (6 mg/L H2O2)
Result for PFBHA derivatization of nonadienal
In order to investigate the reaction mechanism and detect the intermediates or final
products, a higher concentration (10 mg/L) of nonadienal was reacted by UV/H2O2 and then
derivatized with PFBHA. Based on the derivatized chromatograms, there was no difference
between UV photolysis and UV/H2O2 process (Figure 5-7). This result indicates that
nonadienal was removed mainly by UV photolysis in UV/H2O2 process because UV
photolysis is faster than radical reaction and the addition of hydrogen peroxide did not alter
the reaction that produced carbonyls.
77
Figure 5-7. Comparison of PFBHA derivatized chromatograms for UV photolysis and UV/H2O2
treatment of nonadienal
Figure 5-8 shows that carbonyl groups derivatized by PFBHA (oximes) were produced from
the UV irradiation of nonadienal. This result indicates that nonadienal was degraded into
smaller ketone or aldehyde molecules by UV photolysis. Most of these new carbonyl groups
produced from the reaction were not removed by further UV irradiation indicating that these
ketone or aldehyde compounds are highly stable to UV irradiation. However, these ketones or
aldyhydes were not able to be identified in this research. Further study is required to identify
these carbonyl products and to detect other alcoholic or carboxyl products that may be
produced.
78
Figure 5-8. GC/MS chromatograms of PFBHA derivatized nonadienal
Discussion According to the measured molar extinction coefficients and derivatization results,
nonadienal was removed by direct UV photolysis, and a similar mechanism would be
expected for decadienal and heptadienal. While UV photolysis removes fishy/grassy smelling
�dienal� compounds, new types of odors were produced after the oxidation of original
compounds. These transformed odors may be related to carbonyl groups produced from the
UV photolysis of nonadienal, based on the result that these carbonyl groups were not
removed by further UV photolysis. These results are comparable to the results of other
research that reported the fruity smelling aldehydes production from the ozonation (Anselme,
Suffet et al. 1988; AWWARF 1995; Bruchet and Duguet 2004). The C4-C12 normal aldehydes
typically have odor threshold concentrations of < 1µg/L, and are known to be problematic in
drinking water (Fabrellas, Matia et al. 2004). Consequently, carbonyls produced from the
reaction can be one of the causes for the new odors. However, these carbonyl groups
produced by UV photolysis could not be identified and no conclusive evidence was found on
79
the relationship between carbonyl groups produced and new types of odors detected in the
sensory test in this research. Further investigation is required to identify the reaction products,
which may include functional groups other than carbonyls, such as carboxyl or alcohol
groups.
Conclusion The UV/H2O2 process was able to effectively reduce odorous aldehydes concentrations
compared to removal of geosmin and 2-MIB. The result indicates that direct UV photolysis is
the main mechanism involved in this removal. Although the concentration of odorous
aldehydes were reduced by UV/H2O2, new types of odors were produced from these reactions,
which was confirmed by sensory test. Carbonyl groups were detected from the UV photolysis
of nonadienal and were not removed by further UV irradiation. These carbonyl groups were
thought to be related with production of new types of odors such as chalky or sweet odor.
Results indicate that new types of odor can be produced from the oxidation of odorants, and
consequently sensory and chemical analysis should be considered in designing oxidation
process to control recalcitrant odorants.
Acknowledgement
This research was financially supported by Kwater (Korea Water Resources Corporation),
and partially supported by the US National Science Foundation (NSF, Award # 0329474).
The views expressed in this report are those of authors and not those of US NSF.
80
References
Adams, G. E., J. W. Boag, et al. (1965). Absolute Rate constants for the Reaction of the Hydroxyl Radical with Organic Compounds in Pulse Radiolyisis. New York, Academic Press. Andersson, T., G. Forsgren, et al. (2005). "The effects of selected aldehydes, ketones and carboxylic acids on off-flavours in water." International Journal of Food Science and Technology 40(9): 993-1004. Anselme, C., I. H. Suffet, et al. (1988). "Effects of Ozonation on Tastes and Odors." Journal American Water Works Association 80(10): 45-51. AWWA (1991). Guidance Manual for Compliance with the Filtration and Disinfection Requirements for Public Water Systems Using Surface Water. AWWA (1999). Water Quality and Treatment. New York, McGRAW-HILL, Inc. AWWA, APHA, et al. (2005). Standard Method for the Examination of Water and Wastewater. Washington, DC, American Public Health Association. AWWARF (1995). Advances in Taste-and-Odor Treatment and Control. I. H. Suffet, J. Mallevialle and E. Kawczynski, AWWARF. Bader, H. and J. Hoign (1981). "Determination of ozone in water by the indigo method." Water Research 15(4): 449-456. Bao, M. L., F. Pantani, et al. (1998). "Determination of carbonyl compounds in water by derivatization - solid-phase microextraction and gas chromatographic analysis." Journal of Chromatography A 809(1-2): 75-87. Batterman, S., L. Zhang, et al. (2000). "Quenching of chlorination disinfection by-product formation in drinking water by hydrogen peroxide." Water Research 34(5): 1652-1658. Beltran, F. J., G. Ovejero, et al. (1993). "Oxidation of Atrazine in Water by Ultraviolet-Radiation Combined with Hydrogen-Peroxide." Water Research 27(6): 1013-1021. Bennett, J. E. and J. A. Howard (1973). "Bimolecular self-reaction of peroxy radicals. Oxygen-18 isotope study." J. Am. Chem. Soc. 95(12): 4008-4010. Bolton, J. R. and M. I. Stefan (2002). "Fundamental photochemical approach to the concepts of fluence (UV dose) and electrical energy efficiency in photochemical degradation reactions." Research on Chemical Intermediates 28(7-9): 857-870.
81
Bruchet, A. and J. P. Duguet (2004). "Role of oxidants and disinfectants on the removal, masking and generation of tastes and odours." Water Science and Technology 49(9): 297-306. Buchanan, W., F. Roddick, et al. (2006). "Formation of hazardous by-products resulting from the irradiation of natural organic matter: Comparison between UV and VUV irradiation." Chemosphere 63(7): 1130-1141. Burlingame, G. A. and E. D. Mackey (2007). "Philadelphia obtains useful information from its customers about taste and odour quality." Water Science and Technology 55(5): 257-263. Burlingame, G. A., J. J. Muldowney, et al. (1992). "Cucumber Flavor in Philadelphia's Drinking Water." J. American Water Works Association. Butkovskaya, N. I., A. Kukui, et al. (2004). "Rate Constant and Mechanism of the Reaction of OH Radicals with Acetic Acid in the Temperature Range of 229-300 K." J. Phys. Chem. A 108(34): 7021-7026. Carlson, D. A. e. a. (1982). Project Summary: Ultraviolet Disinfection of Water for Small Water Supplies, Environmental Protection Agency. Cassan, D., B. Mercier, et al. (2006). "Effects of medium-pressure UV lamps radiation on water quality in a chlorinated indoor swimming pool." Chemosphere 62(9): 1507-1513. Cater, S. R., M. I. Stefan, et al. (2000). "UV/H2O2 treatment of methyl tert-butyl ether in contaminated waters." Environmental Science & Technology 34(4): 659-662. Chang, P. B. L. and T. M. Young (2000). "Kinetics of methyl tert-butyl ether degradation and by-product formation during UV/hydrogen peroxide water treatment." Water Research 34(8): 2233-2240. Chin, A. and P. R. Berube (2005). "Removal of disinfection by-product precursors with ozone-UV advanced oxidation process." Water Research 39(10): 2136-2144. Cole, S. K., W. J. Cooper, et al. (2007). "Free radical chemistry of disinfection byproducts. 2. Rate constants and degradation mechanisms of trichloronitromethane (chloropicrin)." Environmental Science & Technology 41(3): 863-869. Collivignarelli, C. and S. Sorlini (2004). "AOPs with ozone and UV radiation in drinking water: contaminants removal and effects on disinfection byproducts formation." Water Science and Technology 49(4): 51-56. Collivignarelli, C., Sorlini S. (2004). "AOPs with ozone and UV radiation in drinking water: contaminants removal and effects on disinfection byproducts formation." Water Science and Technology 49(4): 51-56. Cotton, C. A. and J. R. Collins (2006). Dual purpose UV light : Using UV light for Disinfection and for Taste and Odor Oxidation. 2006 WQTC, Denver, CO, USA, AWWA. Cowman, G. A. and P. C. Singer (1996). "Effect of bromide ion on haloacetic acid speciation
82
resulting from chlorination and chloramination of aquatic humic substances." Environmental Science & Technology 30(1): 16-24. Crittenden, J. C., S. M. Hu, et al. (1999). "A kinetic model for H2O2/UV process in a completely mixed batch reactor." Water Research 33(10): 2315-2328. Crittenden, J. C., R. R. Trussell, et al. (2005). Water Treatment: Principles and Design. New Jersey, John Wiley and Sons. Devesa, R., C. Fabrellas, et al. (2004). "The panel of Aigues de Barcelona: 15 years of history." Water Science and Technology 49(9): 145-151. Dietrich, A. M., R. C. Hoen, et al. (1995). "Oxidation of Odorous and Nonodorous Algal Metabolites by Permanganate, Chlorine, and Chlrorine dioxide." Water Science and Technology 31(11): 223-228. Echigo, S., S. Itoh, et al. (2004). "Contribution of brominated organic disinfection by-products to the mutagenicity of drinking water." Water Science and Technology 50(5): 321-328. EPA (1999). EPA Guidance manual : Alternative Disinfectants and Oxidants. Fabrellas, C., L. Matia, et al. (2004). "Determination of odour threshold concentrations and dose-response relations in water of several minor disinfection by-products: aldehydes and alkyl nitriles." Water Science and Technology 49(9): 267-272. Farkas, E., I. Szilagyi, et al. (2003). "Kinetic isotope effect in the reaction of OH radical with acetone-d(6)." Reaction Kinetics and Catalysis Letters 80(2): 351-358. Fliount, R., O. Makogon, et al. (1997). "Radical-mediated degradation mechanisms of tribromo- and other trihalogenated acetic acids in oxygen-free solutions as studied by radiation chemistry methods." Journal of the Chemical Society-Perkin Transactions 2(8): 1535-1545. Glaze W. H., S. R., Chauncey W., Ruth E. C., Zarnoch J. J., Aieta E. M., Tate C. H., and McGuire M.J. (1990). "Evaluating Oxidants for the Removal of Model Taste and Odor Compounds from a Municipal Water Supply." JAWWA 82(5): 79-84. Glaze, W. H., R. Schep, et al. (1990). "Evaluating Oxidants for the Romval of Model Taste and Odor Compound from a Municipal Water-Supply." Journal American Water Works Association 82(5): 79-84. Ho, L., J. P. Croue, et al. (2004). "The effect of water quality and NOM character on the ozonation of MIB and geosmin." Water Science and Technology 49(9): 249-255. Ho, L., G. Newcombe, et al. (2002). "Influence of the character of NOM on the ozonation of MIB and geosmin." Water Research 36: 511-518. Hyun, K. S., Y. J. Kim, et al. (2005). "Profiles of THM and HAAs in the process of Water
83
Treatment Plants." Journal of Korean Society of Water Science and Technology 13(1): 55-66. Jo, C. H., A. M. Dietrich, et al. (2008). Simultaneous Removal of DBPs and Odorants by UV/H2O2 Advanced Oxidation Process Jüttner, F. and S. B. Watson (2007). "Biochemical and Ecological Control of Geosmin and 2-Methylisoborneol in Source Waters." Applied and Environmental Microbiology 73(14): 4395-4406. Khiari, D. (2004). "AWWA's Taste and Odor Committee seeks to understand and solve taste and odor problems in drinking water." Journal American Water Works Association 96(2): 32-+. Klassen, N. V., D. Marchington, et al. (1994). "H2O2 Determination by the I-3(-) Method and by KMNO4 Titration." Analytical Chemistry 66(18): 2921-2925. KMOE (2006). Tasty water guideline report, Korean Ministry of Environment KNIER (2000). Comparison of global drinking water quality standards, Korea National Institute of Environmental Research. Kochany, J. and J. R. Bolton (1992). "Mechanism of Photodegradation of Aqueous Organic Pollutants .2. Measurement of the Primary Rate Constants for Reaction of OH Radicals with Benzene and Some Halobenzenes using an EPR Spin-Trapping Method following the Photolysis of H2O2." Environmental Science & Technology 26(2): 262-265. Krasner, S. W., M. J. McGuire, et al. (1985). "Tastes and Odors - The Flavor Profile Method." Journal American Water Works Association 77(3): 34-39. Krasner, S. W., M. J. McGuire, et al. (1989). "The Occurrence of Disinfection By-products in US Drinking Water." Journal American Water Works Association 81(8): 41-53. Krasner, S. W., H. S. Weinberg, et al. (2006). "Occurrence of a new generation of disinfection byproducts." Environmental Science & Technology 40(23): 7175-7185. Kusakabe, K., S. Aso, et al. (1990). "Decomposition of Humic-Acid and Reductin of Trihalomethane Formation Potential in Water by Ozone with UV Irradiation." Water Research 24(6): 781-785. Lagoa, A. L. C., H. P. Diogo, et al. (2001). "Energetics of C-Cl, C-Br, and C-I bonds in haloacetic acids: Enthalpies of formation of XCH2COOH (X = Cl, Br, I) compounds and the carboxymethyl radical." Chemistry-a European Journal 7(2): 483-489. Lawless, H. T. and H. Heymann (1999). Sensory Evaluation of Food. New York, Chapman & Hall. Lay, Y. S. (1989). Oxidation of 1,2-dibromo-3-chloropropane in ground water using advanced oxidation processes. Los Angeles, University of California. Ph.D.
84
Legrini, O., E. Oliveros, et al. (1993). "Photochemical Processes for Water-Treatment." Chemical Reviews 93(2): 671-698. Li, K., M. I. Stefan, et al. (2004). "UV Photolysis of Trichloroethylene: Product Study and Kinetic Modeling." Environ. Sci. Technol. 38(24): 6685-6693. Li, Y., Y. Xie, et al. (2006). "Photocatalytic hydrogen generation in the presence of chloroacetic acids over Pt/TiO2." Chemosphere 63(8): 1312-1318. Li, Y. L., W. M. Kwok, et al. (2004). "Water-catalyzed dehalogenation reactions of isobromoform and its reaction products." Abstracts of Papers of the American Chemical Society 227: U316-U316. Liang, C. Z., D. S. Wang, et al. (2007). "Kinetics analysis on the ozonation of MIB and geosmin." Ozone-Science & Engineering 29(3): 185-189. Liao, C.-H. and M. D. Gurol (1995). "Chemical Oxidation by Photolytic Decomposition of Hydrogen Peroxide." Environmental Science & Technology 29(12): 3007-3014. Lifongo, L. L., D. J. Bowden, et al. (2004). "Photodegradation of haloacetic acids in water." Chemosphere 55(3): 467-476. Liu, W., S. A. Andrews, et al. (2003). "Optimal methods for quenching H2O2 residuals prior to UFC testing." Water Research 37: 3697-3703. Louis, F., C. A. Gonzalez, et al. (2000). "An ab initio study of the kinetics of the reactions of halomethanes with the hydroxyl radical. 1. CH2Br2." Journal of Physical Chemistry A 104(13): 2931-2938. Louis, F., C. A. Gonzalez, et al. (2000). "An ab initio study of the kinetics of the reactions of halomethanes with the hydroxyl radical. 2. A comparison between theoretical and experimental values of the kinetic parameters for 12 partially halogenated methanes." Journal of Physical Chemistry A 104(38): 8773-8778. Louis, F., C. A. Gonzalez, et al. (2001). "An ab initio study of the kinetics of the reactions of halomethanes with the hydroxyl radical. 3. Kinetic parameters predictions for the potential halon replacements CH2FBr, CHFBr2, CHFClBr, CHCl2Br, and CHClBr2." Journal of Physical Chemistry A 105(9): 1599-1604. Louis, F., C. A. Gonzalez, et al. (2004). "Direct combined ab initio/transition state theory study of the kinetics of the abstraction reactions of halogenated methanes with hydrogen atoms." Journal of Physical Chemistry A 108(47): 10586-10593. Makogon, O., R. Fliount, et al. (1998). "Formation and Degradation of Halogenated Organic Acids. Radiation versus Photocatalytically Induced Processes." Journal of Advanced Oxidation Technology 3(1): 11-21. Mao, Y., C. Schoeneich, et al. (1991). "Identification of organic acids and other intermediates
85
in oxidative degradation of chlorinated ethanes on titania surfaces en route to mineralization: a combined photocatalytic and radiation chemical study." J. Phys. Chem. 95(24): 10080-10089. Martijn, B. J., J. C. Kruithof, et al. (2006). UV/H2O2 treatment : The ultimate solution for organic contaminant control and primary disinfection. Water Quality Technology Conference, Denver, American Water Works Association. Maruthamuthu, P., S. Padmaja, et al. (1995). "Rate Constants for Some Reactions of Free-radicas with Haloacetates in Aqueous-solution." International Journal of Chemical Kinetics 27(6): 605-612. McGivern, W. S., A. Derecskei-Kovacs, et al. (2000). "Computationally efficient methodology to calculate C-H and C-X (X = F, Cl, and Br) bond dissociation energies in haloalkanes." Journal of Physical Chemistry A 104(2): 436-442. McGivern, W. S., A. Derecskei-Kovacs, et al. (2000). "Computationally efficient methodology to calculate C-H and C-X (X = F, Cl, and Br) bond dissociation energies in haloalkanes (vol 104A, pg 436, 2000)." Journal of Physical Chemistry A 104(33): 7916-7916. McGivern, W. S., J. S. Francisco, et al. (2002). "Investigation of the atmospheric oxidation pathways of bromoform: Initiation via OH/Cl reactions." Journal of Physical Chemistry A 106(26): 6395-6400. McGivern, W. S., H. J. Kim, et al. (2004). "Investigation of the atmospheric oxidation pathways of bromoform and dibromomethane: Initiation via UV photolysis and hydrogen abstraction." Journal of Physical Chemistry A 108(35): 7247-7252. Meilgaard, M., G. V. Civille, et al. (1999). Sensory Evaluation Techniques. New York, CRC press. Mezyk, S. P., T. Helgeson, et al. (2006). "Free radical chemistry of disinfection-byproducts. 1. Kinetics of hydrated electron and hydroxyl radical reactions with halonitromethanes in water." Journal of Physical Chemistry A 110(6): 2176-2180. Michael, S. E. and U. Von Gunten (1999). "Hydroxyl Radical/Ozone Ratios During Ozonation Processes. The Rct Concept." Ozone-Science & Engineering 21: 239-260. Milano, J. C., C. Bernatescallon, et al. (1990). "Photolysis of 1,2-dibromo-3-chloropropane in Water." Water Research 24(5): 557-564. Muller, J. P., C. Gottschalk, et al. (2001). "Comparison of advanced oxidation processes in flow-through pilot plants (Part II)." Water Science and Technology 44(5): 311-315. Muller, L. P. and M. Jekel (2001). "Comparison of advanced oxidation processes in flow-through pilot plants (Part I)." Water Science and Technology 44(5): 303-309. Murray, C. A. and S. A. Parsons (2004). "Comparison of AOPs for the removal of natural organic matter: performance and economic assessment." Water Science and Technology
86
49(4): 267-272. Nakatani, N., N. Hashimoto, et al. (2007). "Determination of photoformation rates and scavenging rate constants of hydroxyl radicals in natural waters using an automatic light irradiation and injection system." Analytica Chimica Acta 581(2): 260-267. National Water Research Institute, N. (2000). Treatment Technologies for Removal of Methyl Tertiary Butyl Ether (MTBE) from drinking water. G. Melin. Nicole, I., J. De Laat, et al. (1991). "Degradation of Trihalomethanes in dilute aqueous solution by UV irradiation - Determination of the quantum yield of photolysis at 253.7 nm." Environmental Technology 12: 21-31. Nikolaou, A. D., S. K. Golfinopoulos, et al. (2002). "Determination of haloacetic acids in water by acidic methanol esterification-GC-ECD method." Water Research 36(4): 1089-1094. Oestman, E., L. Schweitzer, et al. (2004). "Effects of chlorine and chloramines on earthy and musty odors in drinking water." Water Science and Technology 49(9): 153-159. Oppenlander, T. (2003). Photochemical Purification of Water and Air, Wiley-VHC. Paradis, N. and A. Hoffman (2006). Mitigation of Taste and Odour Compounds by UV/H2O2 Advanced Oxidation. 2006 WQTC, Denver, CO, USA, AWWA. Parsons, A. F. (2000). An introduction to free radical chemistry. Malden, MA, USA, Blackwell Science Ltd. Paul, H., R. D. Small, et al. (1978). "Hydrogen Abstraction bY Tert-Butoxy radicals - Laser photolysis and Electron-Spin Resonance Study." Journal of the American Chemical Society 100(14): 4520-4527. Pereira, V. J., H. S. Weinberg, et al. (2007). "UV Degradation Kinetics and Modeling of Pharmaceutical Compounds in Laboratory Grade and Surface Water via Direct and Indirect Photolysis at 254 nm." Environ. Sci. Technol. 41(5): 1682-1688. Peter, A. and U. Von Gunten (2007). "Oxidation kinetics of selected taste and odor compounds during ozonation of drinking water." Environmental Science & Technology 41(2): 626-631. Qiao, R. P., N. Li, et al. (2005). "Degradation of microcystin-RR by UV radiation in the presence of hydrogen peroxide." Toxicon 45(6): 745-752. Rahn, R. O. (2004). "Spatial distribution of upper-room germicidal UV radiation as measured with tubular actinometry as compared with spherical actinometry." Photochemistry and Photobiology 80(2): 346-350. Rahn, R. O., J. Bolton, et al. (2006). "The iodide/iodate actinometer in UV disinfection: Determination of the fluence rate distribution in UV reactors." Photochemistry and Photobiology 82(2): 611-615.
87
Rashash, D. M. C., A. M. Dietrich, et al. (1997). "FPA of selected odorous compounds." Journal American Water Works Association 89(4): 131-141. Rashash, D. M. C., A. M. Dietrich, et al. (1995). "The Influence of Growth-conditions on Odor-compound Production by 2 Chrysophytes and 2 Cyanobacteria." Water Science and Technology 31(11): 165-172. Richardson, S. D., M. J. Plewa, et al. (2007). "Occurrence, genotoxicity, and carcinogenicity of regulated and emerging disinfection by-products in drinking water: A review and roadmap for research." Mutation Research-Reviews in Mutation Research 636: 178-242. Richardson, S. D., A. D. Thruston, et al. (2003). "Tribromopyrrole, Brominated Acids, and Other Disinfection Byproducts Produced by Disinfection of Drinking Water Rich in Bromide." Environ. Sci. Technol. 37(17): 3782-3793. Rosenfeldt, E. J. and K. G. Linden (2007). "The R-OH,R-UV concept to characterize and the model UV/H2O2 process in natural waters." Environmental Science & Technology 41(7): 2548-2553. Rosenfeldt, E. J., B. Melcher, et al. (2005). "UV and UV/H2O2 treatment of methylisoborneol (MIB) and geosmin in water." Journal of Water Supply Research and Technology-Aqua 54(7): 423-434. Rosenfeldt, E. J., Melcher Barbara, and Linden, Karl G. (2005). "UV and UV/H2O2 treatment of methylisoborneol(MIB) and geosmin in water." Journal of Water Supply Research and Technology-Aqua 54(7): 423-434. Royce, A. and M. I. Stefan (2005). Application of UV in Drinking Water Treatment for Simultaneous Disinfection and Removal of Taste and Odor Compounds. 2005 WQCT, AWWA. Royce, A. and M. I. Stefan (2005). Application of UV in drinking water treatment for simultaneous disinfection and removal of taste and odor compounds. Rudra, A., N. P. Thacker, et al. (2005). "Hydrogen peroxide and ultraviolet irradiations in water treatment." Environmental Monitoring and Assessment 109(1-3): 189-197. Russell, G. A. (1957). "Deuterium-isotope Effects in the Autoxidation of Aralkyl Hydrocarbons. Mechanism of the Interaction of PEroxy Radicals." J. Am. Chem. Soc. 79(14): 3871-3877. Rychlik, M., P. Schieberle, et al. (1998). Compilation of Odor Threshold, Odor Qualities and Retention Indices of Key Food Odorants. Sarathy, S., Mohseni, M. (2006). "An Overview of UV-based Advanced Oxidation Processes from Drinking Water Treatment." IUVA News 8(2): 16-27. Satchwill, T., S. B. Watson, et al. (2007). "Odourous algal-derived alkenes: differences in
88
stability and treatment responses in drinking water." Water Science and Technology 55(5): 95-102. Sharpless, C. M. and K. G. Linden (2003). "Experimental and model comparisons of low- and medium-pressure Hg lamps for the direct and H2O2 assisted UV photodegradation of N-nitrosodimethylamine in simulated drinking water." Environmental Science & Technology 37(9): 1933-1940. Shemer, H. and K. G. Linden (2006). "Degradation and by-product formation of diazinon in water during UV and UV/H2O2 treatment (vol 136, pg 553, 2003)." Journal of Hazardous Materials 138(3): 638-638. Singer, P. C. (2002). "Occurrence of haloacetic acids in chlorinated drinking water." 3rd World Water Congress: Drinking Water Treatment 2(5-6): 487-492. Song, W. and K. E. O'Shea (2007). "Ultrasonically induced degradation of 2-methylisoborneol and geosmin." Water Research 41: 2672-2678. Spangenberg, D., U. M?ler, et al. (1996). "Photooxidation of exhaust pollutants IV. Photocatalytic and thermal decomposition of trichloroacetic acid: degradation efficiencies and products." Chemosphere 33(1): 43-49. Stefan, M. I. and J. R. Bolton (1998). "Mechanism of the degradation of 1,4-dioxane in dilute aqueous solution using the UV hydrogen peroxide process." Environmental Science & Technology 32(11): 1588-1595. Stefan, M. I., A. R. Hoy, et al. (1996). "Kinetics and mechanism of the degradation and mineralization of acetone in dilute aqueous solution sensitized by the UV photolysis of hydrogen peroxide." Environmental Science & Technology 30(7): 2382-2390. Stefan, M. I., J. Mack, et al. (2000). "Degradation pathways during the treatment of methyl tert-butyl ether by the UV/H2O2 process." Environmental Science & Technology 34(4): 650-658. Suffet, I. H., D. Khiari, et al. (1999). "The drinking water taste and odor wheel for the millennium: Beyond geosmin and 2-methylisoborneol." Water Science and Technology 40(6): 1-13. TrojanUV (2003). Design considerations: 35-43. TrojanUV (2003). Understanging EEO : Water Quality Impacts: 35-43. U.S. EPA (1995). EPA method 552.22. Cincinnati, U.S. Environmental Protection Agency. Vimal, D. and P. S. Stevens (2006). "Experimental and theoretical studies of the kinetics of the reactions of OH radicals with acetic acid, acetic acid-d(3) and acetic acid-d(4) at low pressure." Journal of Physical Chemistry A 110(40): 11509-11516. Vohringer-Martinez, E., B. Hansmann, et al. (2007). "Water Catalysis of a Radical-Molecule
89
Gas-Phase Reaction." Science 315(5811): 497-501. von Gunten, U. and J. Hoigne (1994). "Bromate Formation During Ozonation of Bromide-Containing Waters - Interaction of Ozone and Hydroxyl Radical Reactions." Environmental Science & Technology 28(7): 1234-1242. Wang, G. S., S. T. Hsieh, et al. (2000). "Destruction of humic acid in water by UV light - Catalyzed oxidation with hydrogen peroxide." Water Research 34(15): 3882-3887. Watson, S. B., B. Brownlee, et al. (2000). "Quantitative analysis of trace levels of geosmin and MIB in source and drinking water using headspace SPME." Water Research 34(10): 2818-2828. Watson, S. B., B. Brownlee, et al. (1999). "The use of solid phase microextraction (SPME) to monitor for major organoleptic compounds produced by chrysophytes in surface waters." Water Science and Technology 40(6): 251-256. Watson, S. B., T. Satchwill, et al. (2001). "Under-ice blooms and source-water odour in a nutrient-poor reservoir: biological, ecological and applied perspectives." Freshwater Biology 46(11): 1553-1567. Weinberg, H. S. and W. H. Glaze (1997). "A unified approach to the analysis of polar organic by-products of oxidation in aqueous matrices." Water Research 31(7): 1555-1572. Westerhoff, P., B. Nalinakumari, et al. (2006). "Kinetics of MIB and geosmin oxidation during ozonation." Ozone-Science & Engineering 28(5): 277-286. White, G. C. (1992). Handbook of Chlorination and Alternative Disinfectants. New York, Van Nostrand Reinhold. Wu, C., D. Wei, et al. (2001). "Photosonochemical degradation of trichloroacetic acid in aqueous solution." Chemosphere 44(5): 1293-1297. Xu, B., N. Y. Gao, et al. (2007). "Photochemical degradation of diethyl phthalate with UV/H2O2." Journal of Hazardous Materials 139(1): 132-139. Yokohata, A., T. Ohmura, et al. (1969). "A Dechlorination Reaction in the Radiolysis of Aqueous Monochloroacetic acid Solutions in the presence of Nitrous Oxide." Journal of Physical Chemistry 73(11): 4013-4014. Young, W. F., H. Horth, et al. (1996). "Taste and odour threshold concentrations of potential potable water contaminants." Water Research 30: 331-340. Zalazar, C. S., M. D. Labas, et al. (2007). "Dichloroacetic acid degradation employing hydrogen peroxide and UV radiation." Chemosphere 66(5): 808-815. Zhang, L., W. A. Arnold, et al. (2004). "Kinetics of haloacetic acid reactions with Fe(0)." Environmental Science & Technology 38(24): 6881-6889.
90
Zhao, C. Y., X. F. Lin, et al. (2005). "Water-catalyzed dehalogenation reactions of the isomer of CBr4 and its reaction products and a comparison to analogous reactions of the isomers of di- and trihalomethanes." Chemistry-a European Journal 11(4): 1093-1108.